Metabolic and endometrial ultrasonographic factors associated with clinical pregnancy in PCOS patients undergoing FET: a nomogram development study with internal validation.

Objective: To explore the independent metabolic and ultrasonographic factors associated with clinical pregnancy in patients with polycystic ovary syndrome (PCOS) undergoing frozen-thawed embryo transfer (FET) cycles, develop and validate a nomogram model for predicting the probability of clinical pregnancy in this population, and offer a preliminary framework for risk stratification that requires external validation. Methods: We retrospectively enrolled clinical data of PCOS patients who underwent FET in the Reproductive Medicine Department of General Hospital of Northern Theater Command from January 2017 to September 2022. A total of 2316 eligible FET cycles were included and randomly divided into a training cohort (1621 cycles) and an internal validation cohort (695 cycles) at a ratio of 7:3. All cycles were further divided into clinical pregnancy group and non-clinical pregnancy group according to the clinical pregnancy after transplantation. In the training cohort, the least absolute shrinkage and selection operator (LASSO) regression combined with 10-fold cross-validation was used to screen key predictive variables for clinical pregnancy. Variables with non-zero coefficients were included in multivariate logistic regression analysis to identify independent influencing factors, based on which a nomogram prediction model was constructed. The receiver operating characteristic (ROC) curve, calibration curve, Hosmer-Lemeshow goodness-of-fit test, and decision curve analysis (DCA) were performed to verify the discrimination, calibration, and clinical utility of the model. Results: A total of 2316 FET cycles were included in this study. The clinical pregnancy rate was 63.66% (1032/1621) in the training cohort and 62.16% (432/695) in the internal validation cohort, with no statistically significant difference in baseline data between the two cohorts (P>0.05), indicating good comparability. With lambda.1se as the optimal penalty coefficient, LASSO regression with 10-fold cross-validation finally screened 10 variables with non-zero coefficients that had predictive value for clinical pregnancy, including advanced age, type of infertility, baseline follicle-stimulating hormone (FSH), anti-Müllerian hormone (AMH), total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), fasting plasma glucose (FPG), endometrial thickness, and number of subendometrial blood flow branches. Multivariate logistic regression analysis showed that secondary infertility, number of subendometrial blood flow branches ≥10, and increased endometrial thickness were independent protective factors for clinical pregnancy in PCOS patients undergoing FET cycles, while elevated levels of TC, TG, LDL-C, and FPG were independent risk factors (P<0.05). The nomogram prediction model was constructed based on the above independent influencing factors. The area under the ROC curve (AUC) of the model was 0.74 (95% CI: 0.72~0.77) in the training cohort and 0.75 (95% CI: 0.72~0.79) in the internal validation cohort. The calibration curve showed good consistency between the predicted probability and the actual clinical pregnancy probability, and the Hosmer-Lemeshow goodness-of-fit test indicated no overfitting of the model. DCA results showed that the model could bring favorable clinical net benefit to patients when the threshold probability was in the range of 0.20 to 0.80. Conclusion: This internally validated nomogram, incorporating metabolic and ultrasonographic parameters, demonstrated modest discriminative ability for predicting clinical pregnancy in PCOS patients undergoing FET. However, the model explains only a portion of outcome variation, and external validation with additional predictors (including embryo quality and insulin resistance markers) is needed before clinical application.