Abstract:

External cephalic version (ECV), as an important intervention for correcting breech presentation in the late stage of pregnancy, the prediction of its success rate is crucial for clinical decision-making. Currently, the model-building methods widely used in clinical practice mainly include traditional statistical models and machine-learning models. This paper systematically reviews the application progress of these two types of models in predicting the success rate of ECV. Traditional statistical models, such as the Logistic regression model, are easy to interpret and have strong clinical applicability, but their performance is limited when dealing with complex data. The emerging machine-learning models show better prediction potential, but they also face challenges such as poor interpretability and difficulty in clinical integration. Future research should focus on promoting multicenter, large-sample, and prospective data collection, strengthening the external validation and standardization of models. At the same time, it is necessary to improve the transparency and clinical applicability of machine-learning models, develop prediction tools that can be easily integrated into the clinical process, and ultimately construct a precise and individualized ECV decision-making system to effectively reduce the cesarean section rate without medical indications.

Key words: Breech presentation, Version, fetal, Forecasting, Models, statistical, Machine learning