Predicting the optimal timing for triggering in controlled ovarian stimulation: mature oocytes retrieval predictor

TL;DR

This study develops a model to predict the best time for oocyte retrieval in fertility treatments, improving success rates.

Contribution

A new regression model using FmOI and key clinical variables to optimize ovulation trigger timing in controlled ovarian stimulation.

Findings

The FmOI prediction model achieved high concordance indices (0.98 for follitropin alfa, 0.87 for follitropin delta).

The model reliably predicts mature oocyte counts and improves clinical outcomes in ART.

Lasso regression identified key predictors like initial FSH, follicle count, and gonadotropin dose.

Abstract

The development of assisted reproductive technology (ART) has revolutionized infertility treatment; however, its success largely depends on effective controlled ovarian stimulation (COS) and the timing of oocyte retrieval. This study aimed to develop a regression equation model to optimize the timing of ovulation trigger in COS..

We retrospectively analyzed 503 COS cycles (380 with follitropin alfa, 123 with follitropin delta) as training data. We modified the Follicle-To-Oocyte Index (FOI) and developed the Follicle-To-mature Oocyte Index (FmOI), which indicates how many mature oocytes (MII) were obtained for each antral follicle count. This index was used as an indicator for the retrieval of mature oocytes. When using FmOI as the objective variable, we selected relevant factors through Lasso regression analysis. Based on the obtained regression equations, the accuracy was compared and verified by predicting the number of MII in the test data.

Lasso regression analysis resulted in the creation of an FmOI prediction model using Initial serum FSH, number of follicles ≥ 14 mm, and total gonadotropin dose as explanatory variables. The regression equation model achieved Median Absolute Error values of 1.90 and 1.80 MII counts in the test data for the Alfa and Delta groups, respectively. Concordance index for MII prediction were 0.98 for follitropin alfa and 0.87 for follitropin delta. Use of the model showed higher CLBR in Alfa and non-inferiority in Delta than control group.

This model reliably predicts the number of MII and optimizes trigger timing in COS. By considering key predictors, it provides a precise tool to enhance clinical outcomes in assistedreproductive technology .

Figures