Manually segmented template library for 8-year-old pediatric brain MRI data with 16 subcortical structures.

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Pediatric Neuroscience

Background:

• Manual segmentation of brain MRI is crucial for automated segmentation pipelines.
• Accurate anatomical labeling is essential for pediatric neuroimaging research.

Purpose of the Study:

• To develop a manually delineated pediatric brain MRI template library.
• To evaluate the accuracy of automated segmentation using this pediatric template library.
• To compare segmentation performance with adult templates.

Main Methods:

• Manual segmentation of 16 subcortical structures in 22 pediatric brain MRIs by expert raters.
• Development of a template library from segmented pediatric brain MRI data.
• Cross-validation experiments using randomized template subset selection.
• Automated segmentation pipeline testing and performance evaluation using Dice Similarity Coefficient and Hausdorff distance.

Main Results:

• High accuracy in automated segmentation against manual labels was achieved using the pediatric template library.
• Significant improvement in segmentation accuracy was observed when using age-matched pediatric templates compared to adult templates.
• The developed library provides a benchmark for testing segmentation algorithms.

Conclusions:

• A manually delineated pediatric template library enhances automated brain MRI segmentation accuracy.
• Age-matched templates are critical for precise pediatric neuroimaging analysis.
• This library serves as a valuable resource for developing and validating segmentation algorithms in pediatric populations.