Low field pediatric brain magnetic resonance Image Segmentation and quality Assurance

In early life, accurate description of structural changes in the developing human brain is crucial for understanding healthy development and identification of neurodevelopmental disorders. Existing adult brain deep learning (DL) based segmentation tools struggle with pediatric MRI due to poor gray/white matter differentiation and rapid growth. These challenges impair the ability of existing algorithms to accurately segment pediatric brain structures even with high field (1.5T or 3T) MRI systems. In low and middle-income countries, high field MRI systems are rare due to their high cost and maintenance required. To fill this gap, 0.064T Hyperfine SWOOP scanners are being tested in these settings by the UNITY Consortium, funded by the Bill and Melinda Gates Foundation. Despite lower image quality, low-field MRI offers portability, cost-effectiveness, and eliminates the need for sedation in children. To translate recent improvements in infant brain segmentation to underprivileged communities, we introduce the Low-field pediatric brain magnetic resonance Image Segmentation and quality Assurance (LISA) challenge. We propose the first two tasks of the LISA challenge, to be presented at MICCAI 2024, with additional segmentation tasks to be added as future open challenges. Task 1 involves evaluating objective, quantifiable quality assurance (QA) methods that rate the overall quality of low-field MRI to ensure the acquired MRI meets specific accuracy and consistency standards. Task 2 concerns the automatic segmentation of the bilateral hippocampi, which are pivotal subcortical structures linked to cognitive and memory functions and are often implicated in abnormal neurodevelopment. The overarching goal of this challenge is to develop optimal and publicly available DL tools to assess and segment ultra-low field T2-weighted magnetic resonance images of the brain in early childhood.