Predicting personalized multiple birth risks after in vitro fertilization-double embryo transfer.

Area of Science:

• Reproductive Medicine
• Biostatistics
• Clinical Prediction Modeling

Background:

• In vitro fertilization (IVF) with double embryo transfer (DET) carries significant multiple birth risks.
• Current risk assessment may not fully capture individual patient variability.
• Need for evidence-based, clinic-specific tools for accurate risk prediction.

Purpose of the Study:

• To develop and validate a prediction model (MBP-BIVF) for IVF-DET multiple birth risks.
• To evaluate the model's performance using clinic-specific patient data.
• To assess the utility of personalized risk prediction in patient counseling.

Main Methods:

• Retrospective analysis of 2,413 IVF-DET cycles resulting in live births.
• Utilized boosted tree methods for prediction modeling.
• Validated the model on an independent dataset for predictive power, discrimination, and reclassification.

Main Results:

• Predicted multiple birth probabilities ranged from 11.8% to 54.8%.
• The MBP-BIVF model demonstrated a 146% improvement in predictive power and 16.0% in discrimination over controls.
• Predictions were significantly different from controls in over half of the patients.

Conclusions:

• IVF patients exhibit varying multiple birth risks even with DET, predictable before embryo transfer.
• Clinic-specific prediction models offer an evidence-based, personalized approach to patient counseling.
• Personalized risk prediction enhances informed decision-making in IVF treatment.