Special Issue on MICCAI Open Data 2024-2025

Background. High-quality, openly accessible datasets are essential for developing robust and generalizable medical AI. Yet many diseases, populations, and healthcare settings remain underrepresented in publicly available datasets, limiting the inclusiveness and clinical utility of AI systems.

Initiative. To address this challenge, I co-founded the MICCAI Open Data initiative, a dedicated venue for sharing, discussing, and disseminating medical imaging datasets. Following its launch at MICCAI 2024 in Marrakesh, Morocco, the initiative continued at MICCAI 2025 in Daejeon, South Korea, bringing together researchers committed to open science, FAIR data, and data-centric AI.

Impact. Across the 2024 and 2025 editions, 13 datasets were accepted and collected in this Special Issue of MELBA. The contributions span a diverse range of imaging modalities, diseases, and geographical regions, with a particular emphasis on underrepresented populations and clinical applications.

Vision. The initiative promotes open, reusable, and well-curated datasets as a foundation for trustworthy and reproducible medical AI. By facilitating dataset dissemination and community discussion, we aim to accelerate scientific progress while improving the representativeness of future AI systems.

Role. I served as Guest Editor together with Apostolia Tsirikoglou, Lidia Garrucho Moras, Kaouther Mouheb, and Spyridon Bakas. This Special Issue represents the first major publication output of the MICCAI Open Data initiative and highlights the growing importance of openly accessible data within the medical imaging community.