RISE-MICCAI Journal Club Special Session - LISA Challenge July 11,2025
Thursday 10th July 2025
The RISE-MICCAI Journal Club, in collaboration with the RISE-MICCAI Summer School 2025, is hosting a Special Session featuring the LISA Challenge - Low-field pediatric brain magnetic resonance Image Segmentation and quality Assurance Challenge. In parallel to the traditional version of the Challenge being run as part of MICCAI 2025, the LISA Challenge will run an additional competition solely for participants of the RISE-MICCAI Summer School. Summer School participants will compete in developing deep learning image analysis methods for:
- LISA Task 2: Automated segmentation of the bilateral hippocampus and basal ganglia using ultra-low-field (0.064T) brain MRI.
The top three teams from the Summer School will be awarded prizes and certificates of achievement, and the winners will be featured in a dedicated ranking list on the LISA 2025 website and announced at the MICCAI LISA challenge event.
Friday, July 11, 2025 at 11:00 AM (EDT) / 5:00 PM CET
Abstract
During the initial stages of life, the human brain undergoes rapid tissue growth and development after birth. Accurately capturing and describing structural changes from magnetic resonance images (MRI) by delineating regions of interest (a.k.a segmentation) during this vital period is crucial for gaining new perspectives on healthy brain development and enabling the early identification of neurodevelopmental disorders.
While validated brain segmentation tools exist for adult brains, efforts to directly translate existing adult brain algorithms to pediatric MRI have generally failed. This is partly due to the poor gray/white matter differentiation of the developing brain and its rapid growth throughout the first years of life. 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 the cost and maintenance required. More accessible MRI systems like the 0.064T Hyperfine scanner would help to assess gross anatomical abnormalities; structural delineation challenges are further compounded in these lower resolution imaging environments. Despite the loss in image quality and challenges with image analysis, there are great advantages of using low field MRI, including portability at the point of care of patients, small clinical costs with usability in low resource settings, and elimination of sedation for young children.
To drive advancements in automatic segmentation methods for low-field MRI, RISE-MICCAI summer school participants will be tasked with developing and evaluating deep learning methods for the automatic segmentation of two sets of bilateral subcortical structures, the hippocampi and the basal ganglia, in ultra low field (0.064T) T2-weighted magnetic resonance images of the healthy brain in early childhood (LISA challenge Task 2). The hippocampi are pivotal subcortical structures linked to cognitive and memory functions, and often implicated in abnormal neurodevelopment. The basal ganglia, a group of subcortical nuclei critical for motor control, cognitive functions, and behavioral regulation, are often implicated in disorders involving motor and executive dysfunction.
The LISA challenge provides challenge participants with the first-ever publicly available low-field MRI dataset, ensuring that the training, validation, and testing phases utilize the same dataset obtained from the Synapse platform. The overarching goal of the entire challenge is to obtain optimal deep learning tools for the assessment and segmentation of low-field MRI images in early childhood.