PedAItrics

Artificial intelligence has the potential to transform care delivery in diverse medical specialties by enabling more advanced characterization of biomedical data. However, applications in pediatrics lag behind due to the rare nature of childhood diseases and disorders, the high variability of pediatric patients due to growth and development, as well as unique clinical workflows relative to adults.

Pediatric diseases present distinct challenges including rapid developmental changes, smaller anatomy, motion artifacts, and the limited availability of large, well-annotated datasets. These considerations complicate tasks such as segmentation, registration, and classification, and undermine direct reuse of adult-trained models.

The PedAItrics workshop focuses on the application of medical imaging and artificial intelligence to pediatric, neo/prenatal, and childhood diseases, where imaging plays a central role in diagnosis and longitudinal management. It will showcase recent advances in AI integrating multimodal data — including imaging, digital pathology, -omics, and EHR data — through deep learning and classical machine learning.

By bringing together clinicians, imaging scientists, and AI researchers, PedAItrics aims to identify shared methodological challenges across pediatric diseases and cancers, promote reproducible and trustworthy AI, and accelerate translation into clinical practice.