AutoGlom: Software Tool for Segmentation and Analysis of Magnetic Resonance Images of the Kidney.
Authors
Affiliations (7)
Affiliations (7)
- School of Computing and Augmented Intelligence, Arizona State University, Tempe 85281, Arizona, USA.
- ASU-Mayo Center for Innovative Imaging, Tempe, 85281, Arizona, USA.
- Amazon.
- Meta.
- University of Virginia Children's Hospital, Charlottesville, 22903, Virginia, USA.
- Mallinckrodt Institute of Radiology, Washington University in St. Louis, St. Louis, 63130, Missouri, USA.
- Department of Biomedical Engineering, University of Virginia, Charlottesville VA, USA.
Abstract
Magnetic resonance imaging (MRI) is increasingly important in preclinical and clinical investigations of the kidney. However, there are few user-friendly, flexible, and standardized tools for evaluating MR images for quantitative imaging analysis. Here, we develop AutoGlom, an open-source, modular, and expandable imaging software tool that incorporates artificial intelligence (AI) for segmentation, analysis, and visualization of 3D MR images of the kidney. This initial version of AutoGlom focuses on morphological segmentation and quantification. We describe kidney segmentation from MR images, followed by the use of the graphical user interface of AutoGlom. Using AutoGlom, we measure glomerular number and volume from ex vivo cationic ferritin-enhanced MRI (CFE-MRI) in mice. We further demonstrate a 3D-printed holder to allow for simultaneous imaging of up to 16 mouse kidneys at high resolution (50 µ<i>m</i>) within several hours. The streamlined workflow facilitates rapid image analysis and accelerates optimization of cationic ferritin dosing and imaging parameters. These tools are a resource for the kidney community that may accelerate the identification of candidate imaging biomarkers from 3D MRI of the kidney and have the potential to be extended to in vivo studies and other imaging modalities.