ATOMMIC: An Advanced Toolbox for Multitask Medical Imaging Consistency to facilitate Artificial Intelligence applications from acquisition to analysis in Magnetic Resonance Imaging

Dimitrios Karkalousos,Ivana Išgum,Henk A. Marquering,Matthan W.A. Caan

DOI: https://doi.org/10.2139/ssrn.4801289

2024-05-01

Abstract:AI is revolutionizing MRI along the acquisition and processing chain. Advanced AI frameworks have been developed to apply AI in various successive tasks, such as image reconstruction, quantitative parameter map estimation, and image segmentation. Existing frameworks are often designed to perform tasks independently or are focused on specific models or datasets, limiting generalization. We introduce ATOMMIC, an open-source toolbox that streamlines AI applications for accelerated MRI reconstruction and analysis. ATOMMIC implements several tasks using DL networks and enables MultiTask Learning (MTL) to perform related tasks integrated, targeting generalization in the MRI domain. We first review the current state of AI frameworks for MRI through a comprehensive literature search and by parsing 12,479 GitHub repositories. We benchmark 25 DL models on eight publicly available datasets to present distinct applications of ATOMMIC on accelerated MRI reconstruction, image segmentation, quantitative parameter map estimation, and joint accelerated MRI reconstruction and image segmentation utilizing MTL. Our findings demonstrate that ATOMMIC is the only MTL framework with harmonized complex-valued and real-valued data support. Evaluations on single tasks show that physics-based models, which enforce data consistency by leveraging the physical properties of MRI, outperform other models in reconstructing highly accelerated acquisitions. Physics-based models that produce high reconstruction quality can accurately estimate quantitative parameter maps. When high-performing reconstruction models are combined with robust segmentation networks utilizing MTL, performance is improved in both tasks. ATOMMIC facilitates MRI reconstruction and analysis by standardizing workflows, enhancing data interoperability, integrating unique features like MTL, and effectively benchmarking DL models.

Medical Physics,Artificial Intelligence,Software Engineering,Mathematical Physics

What problem does this paper attempt to address?

The paper attempts to address the limitations of existing artificial intelligence (AI) frameworks for magnetic resonance imaging (MRI) when performing multiple tasks. Specifically, existing frameworks are often designed to perform specific tasks independently or focus on specific models or single datasets, which limits their generalization capabilities in the MRI field. To address these issues, the authors introduce an open-source toolbox named ATOMMIC (Advanced Toolbox for Multitask Medical Imaging Consistency), which aims to improve the performance of MRI image reconstruction and analysis by integrating related tasks into a unified framework through multitask learning (MTL) techniques.

### Main Issues
1. **Task Independence**: Existing frameworks are typically designed to perform specific tasks independently, such as image reconstruction, quantitative parameter map estimation, and image segmentation, lacking support for the integration of multiple tasks.
2. **Data Type Support**: Existing frameworks exhibit inconsistencies in supporting complex and real domain data, limiting their application across different tasks.
3. **Generalization Capability**: Existing frameworks are often optimized for specific datasets or models, lacking generalization capability across different MRI tasks.

### Solutions
1. **Multitask Learning (MTL)**: ATOMMIC integrates related tasks (such as image reconstruction and image segmentation) into a unified framework through multitask learning techniques, improving task performance and efficiency.
2. **Data Type Support**: ATOMMIC uniformly supports complex and real domain data, ensuring data consistency across different tasks.
3. **Standardized Workflow**: ATOMMIC provides a standardized workflow, enhancing data interoperability and effectively benchmarking deep learning models.
4. **Comprehensive Documentation and Support**: ATOMMIC offers detailed documentation and ongoing maintenance, facilitating researchers in using and developing new AI applications.

### Objectives
- **Improve MRI Image Reconstruction and Analysis Performance**: Enhance the accuracy of image reconstruction and analysis through multitask learning and data consistency support.
- **Provide a Comprehensive AI Framework**: Offer researchers a comprehensive AI framework that can be used for various MRI tasks, promoting the development and sharing of new AI applications.
- **Enhance Generalization Capability**: Improve the framework's generalization capability across different MRI tasks by supporting multiple data types and tasks.

In summary, this paper aims to address the shortcomings of existing AI frameworks in terms of multitask support, data type support, and generalization capability by introducing the ATOMMIC toolbox, thereby advancing research and applications in the MRI field.