Multi-Institutional Annotated Multiparametric MRI Dataset of Pediatric High-Grade Gliomas.

Area of Science:

• Neuro-oncology
• Medical Imaging
• Artificial Intelligence

Background:

• Pediatric brain tumors are a leading cause of cancer mortality in children, presenting unique challenges compared to adult tumors.
• Progress in pediatric neuro-oncology AI is limited by a lack of large, standardized datasets.
• Existing datasets do not adequately support the development and validation of AI algorithms for pediatric brain tumor analysis.

Purpose of the Study:

• To introduce the Brain Tumor Segmentation in Pediatrics (BraTS-PEDs) dataset, the first large-scale, open-access benchmark for pediatric brain tumor segmentation.
• To provide a standardized resource for developing and evaluating AI algorithms in pediatric neuro-oncology.
• To facilitate reproducible research and model generalization across institutions.

Main Methods:

• The BraTS-PEDs dataset includes multiparametric MRI scans from 457 pediatric patients with high-grade gliomas.
• Scans comprise pre- and post-contrast T1-weighted, T2-weighted, and T2-FLAIR sequences.
• Tumor subregions were annotated using a semi-automated process combining auto-segmentation and expert manual refinement, following RAPNO recommendations.

Main Results:

• The dataset is partitioned into training (n=257), validation (n=91), and hidden testing (n=109) subsets for reproducible benchmarking.
• BraTS-PEDs is the first large-scale, standardized resource for pediatric brain tumor segmentation and analysis.
• It enables the development and evaluation of AI algorithms tailored for pediatric neuro-oncology.

Conclusions:

• The BraTS-PEDs dataset is a foundational resource for advancing AI in pediatric neuro-oncology.
• It supports reproducible method comparison and model generalization.
• Facilitates future integration of imaging with molecular and clinical data for precision medicine in pediatric brain tumors.