Related Experiment Video
Updated: Sep 13, 2025
06:38
In Vitro Modeling of Down Syndrome Neurogenesis Using Human-Induced Pluripotent Stem Cells
Published on: March 7, 2025
621
Ensemble learning-driven hybrid prediction model for improved prenatal down's syndrome screening: a comparative study
Liping Hu1, Minglin Xu1, Yuqin Li2
1Department of Laboratory Medicine, Xiamen Chang Gung Hospital Hua Qiao University, Xiamen 361028, PR China.
Summary
A new hybrid prediction model (HyPred) using ensemble learning (EL) significantly improves prenatal Down
Area of Science:
- Genetics and Genomics
- Computational Biology
- Maternal-Fetal Medicine
Background:
- Prenatal screening for Down's syndrome (DS) is crucial for early detection and management.
- Existing screening methods, such as laboratory-based median equations, have limitations in accuracy and adaptability.
- The integration of advanced computational techniques offers potential for enhanced predictive performance.
Purpose of the Study:
- To develop and validate a hybrid prediction model (HyPred) for prenatal Down's syndrome (DS) screening.
- To leverage ensemble learning (EL) techniques, including extreme gradient boosting, balanced random forest, and gradient boosting machine, for improved prediction.
- To compare the performance of HyPred against traditional laboratory-based median equations.
Main Methods:
- Retrospective analysis of 8,363 first-trimester samples for model training (Nov 2019-Aug 2022).
- Validation using 1,943 independent first-trimester samples (Sep 2022-Jul 2023).
- Development of the HyPred model in R, comparing its sensitivity, specificity, and other performance metrics against a laboratory-based median equation.
Main Results:
- The laboratory-based median equation demonstrated improved screening performance compared to the default equation.
- The ensemble learning (EL) model, HyPred, exhibited superior performance in the validation set, showing higher accuracy, robustness, and adaptability.
- HyPred achieved an Area Under the Curve (AUC) of 0.97, outperforming the median equation across key performance metrics.
Conclusions:
- The ensemble learning (EL) model provides enhanced accuracy and robustness for prenatal Down's syndrome (DS) screening.
- While computationally intensive, modern tools enable optimization of EL models for clinical application.
- This approach supports precision medicine by enabling more customized and effective prenatal screening strategies.
