Development and validation of a nomogram integrating endometrial ultrasound parameters to predict clinical pregnancy in frozen-thawed embryo transfer cycles.

OBJECTIVE

To develop and validate a multivariable logistic regression prediction model integrating ultrasound-based determinants influencing clinical pregnancy outcomes following frozen-thawed embryo transfer (FET) and to establish a clinically applicable prediction model integrating morphological and functional endometrial parameters.

METHODS

This study conducted uterine endometrium ultrasound evaluations on 325 infertile patients before transplantation. According to the inclusion and exclusion criteria, 107 infertile women were excluded, and 218 patients were included in the subsequent analysis. To ensure robust model evaluation and adhere to TRIPOD guidelines, the 218 participants were randomly split into a training cohort (n=154, ~70%) for model development and an validation cohort (n=64, ~30%) for performance assessment. Baseline characteristics, hormone levels, and ultrasound parameters were compared between the two groups. Univariate analysis, collinearity diagnosis, and multivariate logistic regression was utilized to determine independent predictors and construct a nomogram. A nomogram model was constructed in the training cohort and validated in the independent validation cohort, supplemented by1,000 bootstrap iterations. The performance of the model was assessed through discrimination, calibration, decision curve analysis (DCA), and clinical impact curves (CIC).

RESULTS

Multivariate analysis identified endometrial thickness (ET; OR = 1.315, P = 0.037) on the day before transfer, resistance index (RI; OR = 0.014, P = 0.027), systolic velocity/diastolic velocity (S/D; OR = 0.531, P = 0.047), uterine peristalsis (UP) direction (OR = 0.598, P = 0.022), and UP frequency (OR = 0.653, P = 0.039) as independent predictors of pregnancy outcomes. The nomogram demonstrated good discrimination (AUC = 0.828) and satisfactory calibration (Hosmer-Lemeshow P = 0.91). DCA shows that the model has a net clinical benefit in the probability range of 10% to 80%, and the CIC plot confirmed excellent concordance between predicted and observed pregnancy probability.

CONCLUSION

The developed nomogram exhibits strong performance and clinical utility in predicting the possibility of pregnancy following FET, providing a non-invasive and practical tool for individualized clinical decision-making.