Uncertainty in clinical data and stochastic model for in vitro fertilization

Area of Science:

• Reproductive Medicine
• Biomedical Engineering
• Mathematical Modeling

Background:

• In vitro fertilization (IVF) is a key assisted reproductive technology (ART) involving superovulation, egg retrieval, fertilization, and embryo transfer.
• Superovulation aims for multiple follicle growth using drug-induced methods, but is subject to patient-specific factors and inherent uncertainties.
• Clinical data in IVF superovulation is affected by measurement noise, necessitating robust models for follicle growth prediction.

Purpose of the Study:

• To develop a robust stochastic model for projecting superovulation cycle outcomes in IVF.
• To account for process noise and patient response variability in modeling follicle growth.
• To improve the accuracy of predicting IVF treatment success by addressing uncertainties.

Main Methods:

• Development of a stochastic model to represent follicle growth dynamics during superovulation.
• Comparison of stochastic model predictions against clinical data and a deterministic model.
• Evaluation of prediction accuracy and trend matching for different patient responses.

Main Results:

• The stochastic model demonstrated improved projection accuracy for superovulation cycle outcomes compared to deterministic models.
• Prediction accuracy was enhanced by over 70% for specific patients using the stochastic approach.
• The stochastic model provided a better match with clinical data trends, even when significant accuracy improvements were not observed.

Conclusions:

• Stochastic modeling offers a more robust approach to predicting IVF superovulation outcomes by incorporating process noise.
• This enhanced modeling can lead to more accurate and reliable predictions, potentially improving IVF treatment planning.
• The findings highlight the importance of accounting for uncertainty in biological processes for developing effective clinical models.