Predictive models for live birth outcomes following fresh embryo transfer in assisted reproductive technologies using...

Area of Science:

• Reproductive Medicine
• Artificial Intelligence in Healthcare
• Biostatistics

Background:

• Infertility affects 15% of global couples, with assisted reproductive technologies (ARTs) as primary interventions.
• ART success rates have plateaued around 30%, necessitating improved predictive models for better outcomes.
• This study focuses on developing a machine learning-based predictive model for live birth after fresh embryo transfer.

Purpose of the Study:

• To develop and evaluate machine learning models for predicting live birth outcomes in fresh embryo transfer.
• To identify key pre-pregnancy features influencing live birth rates in ART.
• To create a tool for clinicians to aid in treatment individualization and patient counseling.

Main Methods:

• Analysis of 11,728 ART records (2016-2023) with 55 pre-pregnancy features.
• Implementation and comparison of six machine learning models: Random Forest (RF), XGBoost, GBM, AdaBoost, LightGBM, and ANN.
• Validation of model performance using area under the curve (AUC).

Main Results:

• Random Forest (RF) achieved the highest predictive performance with an AUC > 0.8.
• Key predictors identified include female age, embryo grade, number of usable embryos, and endometrial thickness.
• A web-based tool was developed for clinical prediction and treatment individualization.

Conclusions:

• Machine learning models significantly advance the prediction of live birth outcomes before embryo transfer.
• The study highlights the potential of AI to enhance clinical decision-making and patient counseling in ART.
• This approach moves beyond traditional assessments, offering a more personalized strategy for infertility treatment.