逻辑回归
超声波
子宫内膜
医学
接收机工作特性
体质指数
妇科
产科
放射科
内科学
作者
Meiling Li,Xianjun Zhu,Liping Wang,Haiyan Fu,Zhao Wei,Chen Zhou,Li Chen,Bing Yao
标识
DOI:10.1080/02648725.2023.2183585
摘要
Up to today, there is no effective, specific and non-invasive evaluation method to assess the endometrial receptivity. This study aimed to establish a non-invasive and effective model with the clinical indicators to evaluate endometrial receptivity. Ultrasound elastography can reflect the overall state of the endometrium. Ultrasonic elastography images from 78 hormonally prepared frozen embryo transfer (FET) patients were assessed in this study. Meanwhile, the clinical indicators reflecting endometrium in the transplantation cycle were collected. The patients were received to transfer only one high-quality blastocyst. A novel code rule that can generate a large number of 0-1 symbols was designed to collect data on different factors. At the same time, a logistic regression model of the machine learning process with an automatic combination of factors was designed for analysis. The logistic regression model was based on age, body mass index, waist-hip ratio, endometrial thickness, perfusion index (PI), resistance index (RI), elastic grade, elastic ratio cutoff value, serum estradiol level and 9 other indicators. The accuracy rate of predicting pregnancy outcome of the logistic regression model was 76.92%. Elastic ultrasound can reflect the endometrial receptivity of patients in FET cycles. We established a prediction model including ultrasound elastography and the model precisely predicted the pregnancy outcome. The predictive accuracy of endometrial receptivity by the predictive model is significantly higher than that of the single clinical indicator. The prediction model by integrating the clinical indicators to evaluate endometrial receptivity may be a non-invasive and worthwhile method for evaluating endometrial receptivity.
科研通智能强力驱动
Strongly Powered by AbleSci AI