Analysis of Influencing Factors of Embryo Development Arrest During Early Pregnancy and Construction and Validation of its Prediction Model.

ABSTRACT

Context The early symptoms of embryo development arrest are not typical. There is currently no model tool available to predict embryo development arrest. Objective: To explore the influencing factors of embryo development arrest in early pregnancy and build a risk prediction model. Methods: From May 2019 to March 2023, 277 patients suspected of embryonic development arrest during the first ultrasound examination in the Department of Obstetrics and Gynecology of the Ninth Affiliated Hospital of Soochow University were retrospectively selected as the study subjects. They were divided into diapause group and non-diapause group according to the second ultrasound (review after 1-2 weeks) to diagnose whether embryo development arrest. Collect two sets of data for analysis, Screen out the influencing factors of early pregnancy embryo development arrest. The logistic regression model and random forest model were constructed respectively. Evaluate the predictive performance of two statistical models. Results: Out of 277 suspected cases of embryonic developmental arrest, 88 were ultimately confirmed. Older age (OR 2.259, P = .017), higher ultrasonic blood flow resistance index (RI) (OR 1.728, P = .038), higher ultrasonic gestational sac diameter/embryo head hip length ratio (MSD/CRL) (OR1.919, P = .007), lower progesterone (OR 0.562, P = .011), and lower pregnancy-associated protein A (PAPP-A) (OR 0.495, P = .023). The low expression of vascular endothelial growth factor (VEGF) (OR 0.618, P = .005) was the influencing factor of embryo development arrest in early pregnancy. Building a prediction model based on the above indicators, it was found through testing that the random forest model is superior to the logistic regression model in predicting the risk of embryo development arrest. Conclusion: A random forest model based on age, ultrasound RI, progesterone, PAPP-A, ultrasound MSD/CRL ratio, and VEGF index can help clinicians identify the risk of embryonic developmental arrest.