Swiss researchers developed an AI model that automatically segments major anatomic structures in MRI scans, outperforming publicly available tools in accuracy and consistency.
Summary:
Swiss researchers have developed an AI model that automatically segments major anatomic structures in MRI images, regardless of the imaging sequence. The open-source tool, based on the nnU-Net framework, was trained using a large dataset of MRI and CT scans and demonstrated superior performance compared to publicly available segmentation tools. In a retrospective study, the model achieved high accuracy, with a Dice score of 0.839, and performed well across different MRI scanners and imaging protocols. Researchers suggest the AI tool could enhance radiology workflows, improve measurement precision, and be used for treatment planning, disease monitoring, and opportunistic screening in clinical settings.
Key Takeaways:
- AI Model Improves MRI Segmentation: The AI-powered TotalSegmentator MRI automates segmentation of 80 anatomic structures, reducing radiologists’ workload and improving consistency.
- High Accuracy and Versatility: The model achieved a Dice score of 0.839 and performed well across different MRI scanners and imaging sequences, surpassing existing segmentation tools.
- Potential Clinical Applications: Researchers suggest the AI tool could assist in treatment planning, disease monitoring, and opportunistic screening, making MRI-based diagnostics more efficient.
Research scientists in Switzerland have developed and tested a robust artificial intelligence (AI) model that automatically segments major anatomic structures in MRI images, independent of sequence, according to a new study published in Radiology. In the study, the model outperformed other publicly available tools.
For in-depth interpretation of MRI images, the organs, muscles, and bones in the images are outlined or marked, which is known as segmenting.
“MRI images have traditionally been manually segmented, which is a time-consuming process that requires intensive effort by radiologists and is subject to inter-reader variability,” says Jakob Wasserthal, PhD, radiology department research scientist at University Hospital Basel in Basel, Switzerland, in a release. “Automated systems can potentially reduce radiologist’s workload, minimize human errors and provide more consistent and reproducible results.”
Developing the TotalSegmentator MRI Model
Wasserthal and colleagues built an open-source automated segmentation tool called the TotalSegmentator MRI based on nnU-Net, a self-configuring framework that has set new standards in medical image segmentation. It adapts to any new dataset with minimal user intervention, automatically adjusting its architecture, preprocessing, and training strategies to optimize performance. A similar model for CT (TotalSegmentator CT) is being used by over 300,000 users worldwide to process over 100,000 CT images daily.
In the retrospective study, the researchers trained TotalSegmentator MRI to provide sequence-independent segmentations of major anatomic structures using a randomly sampled dataset of 616 MRI and 527 CT exams.
The training set included segmentations of 80 anatomic structures typically used for measuring volume, characterizing disease, surgical planning, and opportunistic screening.
“Our innovation was creating a large data set,” Wasserthal says in a release. “We used a lot more data and segmented many more organs, bones, and muscles than has been previously done. Our model also works across different MRI scanners and image acquisition settings.”
To evaluate the model’s performance, Dice scores—which measure how similar two sets of data are—were calculated between predicted segmentations and radiologist reference standards for segmentations. The model performed well across the 80 structures with a Dice score of 0.839 on an internal MRI test set. It also significantly outperformed two publicly available segmentation models (0.862 versus 0.838 and 0.560) and matched the performance of TotalSegmentator CT.
Potential for Clinical Applications
“To our knowledge, our model is the only one that can automatically segment the highest number of structures on MRIs of any sequence,” he says in a release. “It’s a tool that helps improve radiologists’ work, makes measurements more precise, and enables other measurements to be done that would have taken too much time to do manually.”
In addition to research and AI product development, Wasserthal says the model could potentially be used clinically for treatment planning, monitoring disease progression, and opportunistic screening.
Photo caption: Axial MRI images from the cases with the lowest (top) and highest (bottom) Dice score in the CHAOS external test set for our proposed model, TotalSegmentator MRI, as well as for two publicly available baseline models, MRSegmentator and AMOS. The reference segmentation for liver and spleen is shown in green, and the segmentation of the model is shown in red. The CHAOS dataset was used to show the best and the worst results because this dataset is the most independent from the training data of the three models.
Photo credit: Radiological Society of North America
