Prediction Model of Late Fetal Growth Restriction with Machine Learning Algorithms.

Autor:	Lee SU; Department of Obstetrics and Gynecology, Incheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea., Choi SK; Department of Obstetrics and Gynecology, Incheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea., Jo YS; Department of Obstetrics and Gynecology, St. Vincent's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea., Wie JH; Department of Obstetrics and Gynecology, Eunpyeong St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea., Shin JE; Department of Obstetrics and Gynecology, Bucheon St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea., Kim YH; Department of Obstetrics and Gynecology, Uijeongbu St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea., Kil K; Department of Obstetrics and Gynecology, Yeouido St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea., Ko HS; Department of Obstetrics and Gynecology, Seoul St. Mary's Hospital, College of Medicine, The Catholic University of Korea, Seoul 06591, Republic of Korea.
Jazyk:	angličtina
Zdroj:	Life (Basel, Switzerland) [Life (Basel)] 2024 Nov 20; Vol. 14 (11). Date of Electronic Publication: 2024 Nov 20.
DOI:	10.3390/life14111521
Abstrakt:	Background: This study aimed to develop a clinical model to predict late-onset fetal growth restriction (FGR). Methods: This retrospective study included seven hospitals and was conducted between January 2009 and December 2020. Two sets of variables from the first trimester until 13 weeks (E1) and the early third trimester until 28 weeks (T1) were used to develop the FGR prediction models using a machine learning algorithm. The dataset was randomly divided into training and test sets (7:3 ratio). A simplified prediction model using variables with XGBoost's embedded feature selection was developed and validated. Results: Precisely 32,301 patients met the eligibility criteria. In the prediction model for the whole cohort, the area under the curve (AUC) was 0.73 at E1 and 0.78 at T1 and the area under the precision-recall curve (AUPR) was 0.23 at E1 and 0.31 at T1 in the training set, while an AUC of 0.62 at E1 and 0.73 at T1 and an AUPR if 0.13 at E1, and 0.24 at T1 were obtained in the test set. The simplified prediction model performed similarly to the original model. Conclusions: A simplified machine learning model for predicting late FGR may be useful for evaluating individual risks in the early third trimester. Competing Interests: The authors declare no conflicts of interest.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu Plný text ve formátu PDF