Technical Note: MRQy - An open-source tool for quality control of MR imaging data.

Area of Science:

• Medical Imaging
• Computational Biology
• Data Science

Background:

• Large imaging datasets are crucial for developing computational models in medical research.
• Magnetic resonance imaging (MRI) datasets can exhibit significant variations in quality, impacting model generalizability.
• Standardized quality control is needed to address variations in MRI resolution, contrast, and artifacts.

Purpose of the Study:

• To introduce MRQy, a quantitative quality control tool for MRI datasets.
• To enable rapid assessment of site- or scanner-specific variations in image resolution and contrast.
• To identify imaging artifacts like noise and inhomogeneity for correction before model development.

Main Methods:

• MRQy extracts quality measures (e.g., noise ratios, variation metrics) and metadata (e.g., voxel resolution) from any body region's MRI.
• A generalized approach allows MRQy to work across diverse imaging cohorts.
• An HTML5 front-end provides real-time filtering and trend visualization of quality measures.

Main Results:

• MRQy successfully identified site-specific variations in brain and rectal MRI cohorts, indicating potential batch effects.
• Unsupervised clustering using MRQy measures accurately distinguished between imaging sites before processing.
• After processing with MRQy, batch effects were largely mitigated, and outlier datasets with artifacts were identified.

Conclusions:

• MRQy is a standalone, unsupervised tool for efficient MRI quality control on standard computers.
• The tool is freely accessible and open-source, promoting community use and feedback.
• MRQy aids in standardizing MRI data, improving the reliability of computational models.