The RSNA has bestowed its 2025 Alexander R. Margulis Award on a study showing that an open-source AI model can automatically segment anatomic structures on MR images independent of sequence.
Led by researchers at the University Hospital Basel in Switzerland, the study described the development and testing of TotalSegmentator MRI, a model based on a self-configuring framework called nnU-Net, a type of AI model designed to automatically recognize and outline structures in images.
The researchers trained the model to analyze age-related changes in organ volumes on a diverse set of major anatomic structures using 616 MRI and 527 CT images and then tested it on a large internal dataset of 8,672 abdominal MRI scans. They observed expected patterns, such as declining kidney, liver, and spleen volumes with age, and increasing adrenal gland volumes.
The study was published on February 18 in Radiology. Lead author Tugba Akinci D’Antonoli, MD, will accept the award at the upcoming RSNA meeting.
