Fraunhofer MEVIS successfully contributed to the 28th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) held from September 23 to 27, 2025 in Daejeon, Republic of Korea. The annual MICCAI conference is a forum for world leading biomedical scientists, engineers, and clinicians from a wide range of disciplines associated with medical imaging and computer assisted intervention.

Researchers of Fraunhofer MEVIS have excelled in the following areas:

• UNICORN Lighthouse Challenge to assess the performance of multimodal foundation models in medical imaging across multiple clinical tasks: The Fraunhofer MEVIS team with Raphael Schäfer, Till Nicke, Habib Mergan, Fabian Kiessling, and Johannes Lotz won the overall rating as well as in the categories pathology and radiology. The TissueConcepts / M3 model was the only universal model that did not need to be specialized towards one of the modalities. Medical foundation models are AI base models that can be adapted to new tasks using small datasets. Thanks to their broad pre-training, foundation models increase robustness also in the derived models. In contrast to the common self-supervised training, our supervised training approach uses only a fraction of the energy with respect to training cost and CO₂ emissions.
  
  
• MAMA-MIA Challenge to advance generalizability and fairness in breast MRI tumor segmentation and treatment response prediction: The Fraunhofer MEVIS team with Kai Geissler and Raphael Schäfer won second position out of 20 participants on Task 1 (Primary Tumor Segmentation on Breast Dynamic Contrast-Enhanced MRI) and second position out of 15 participants on Task 2 (Prediction of Pathologic Complete Response to Neoadjuvant Chemotherapy). Breast cancer remains the most common cancer among women and a leading cause of female mortality. Dynamic contrast-enhanced MRI (DCE-MRI) is a powerful imaging tool for evaluating breast tumors, yet the field lacks a standardized benchmark for analyzing treatment responses and guiding personalized care.
  
  
• CARE-WHS Challenge on Whole Heart Segmentation (WHS): The Fraunhofer MEVIS researchers Lisa Bautz and Johanna Brosig won third place out of 10 participants. Cardiovascular diseases, as the leading cause of death globally, necessitate precise morphological and pathological quantification through segmentation of crucial cardiac structures from medical images. WHS faces challenges including heart shape variability during the cardiac cycle, clinical artifacts like motion and poor contrast-to-noise ratio, as well as domain shifts in multi-center data and the distinct modalities of CT and MRI. The CARE-WHS challenge serves to inspire innovative solutions for these tasks.
  
  
• Poster Award in the Workshop Fairness of AI in Medical Imaging (FAIMI): Dishantkumar Sutariya and Eike Petersen were awarded for their poster and paper meval: A Statistical Toolbox for Fine-Grained Model Performance Analysis.

In addition, Fraunhofer MEVIS researchers

• Tanja Loßau, Jan Hendrik Moltz, and Temke Kohlbrandt co-organized the Automated Lesion Segmentation in Whole-Body PET/CT Challenge,
• Markus Wenzel and Lars Ole Schwen co-organized the ODELIA Breast MRI Challenge, and
• Eike Petersen co-organized the FAIMI workshop and gave the FAIMI tutorial lecture.