Related Experiment Video
Updated: Oct 3, 2025
04:25
Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
Published on: December 15, 2023
3.0K
Automated 3D Fetal Brain Segmentation Using an Optimized Deep Learning Approach
L Zhao1,2, J D Asis-Cruz1, X Feng3
1From the Department of Diagnostic Imaging and Radiology (L.Z., J.D.A.-C., Y.W., K.K., A.L., J.Q., C. Lopez, C. Limperopoulos), Developing Brain Institute, Children's National, Washington, DC.
AJNR. American Journal of Neuroradiology
|February 18, 2022
Summary
A new deep learning method automates fetal brain segmentation from MR imaging, offering improved accuracy and reliability over manual and atlas-based techniques. This advance enhances the study of fetal brain development and disease.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Neuroscience
Background:
- Magnetic resonance (MR) imaging is crucial for assessing fetal brain development.
- Current manual segmentation methods are time-consuming and lack repeatability.
- Existing atlas-based methods have limitations in accuracy and robustness.
Purpose of the Study:
- To develop a deep learning-based automatic fetal brain segmentation method.
- To improve accuracy and robustness compared to traditional methods.
- To provide a reliable tool for clinical and research applications.
Main Methods:
- Trained a deep learning model on 65 fetal MR imaging studies (23-39 weeks gestation).
- Compared the model's performance against a 4D atlas-based segmentation method.
- Evaluated the model on 41 fetuses with congenital heart disease.
Main Results:
- Achieved high consistency with manual segmentation (average Dice score of 0.897).
- Demonstrated significantly improved performance over atlas-based methods (P < .001).
- Showed consistent performance across gestational ages and in fetuses with congenital heart disease.
Conclusions:
- The deep learning method offers an efficient and reliable approach for fetal brain segmentation.
- Outperformed 4D atlas-based segmentation, showing clinical and research utility.
- Provides a robust tool for analyzing fetal brain growth and development.
