Predicting live birth by combining cleavage and blastocyst-stage time-lapse variables using a hierarchical and a data...

Area of Science:

• Reproductive Medicine
• Embryology
• Biostatistics

Background:

• Prolonged embryo culture and single embryo transfer are key for improving pregnancy rates.
• Limited research exists on the predictive value of time-lapse monitoring (TLM) parameters from later embryo developmental stages (morula and blastocyst).

Purpose of the Study:

• To evaluate the predictive capacity of TLM parameters from later embryo developmental stages for live birth in single vitrified-thawed blastocyst transfers (SVBT).

Main Methods:

• Retrospective analysis of 285 SVBT cases using TLM data.
• Development of two statistical models: a hierarchical model with key predictors (t2, texpB2) and a data-mining model using principal component analysis (PCA).

Main Results:

• The hierarchical model stratified blastocysts into four categories with increasing live birth (LB) rates (8% to 53%).
• The data-mining model showed increasing LB rates across quartiles of the Sc parameter (12% to 49%).
• Both models demonstrated comparable predictive accuracy (AUC ~0.72).

Conclusions:

• Combining cleavage- and blastocyst-stage TLM variables successfully predicts live birth potential.
• Further internal and external validation is needed to confirm the generalizability of these predictive models.