mIQa: A Highly Scalable and Customizable Platform for Medical Image Quality Assessment - Phase II

点击翻译按钮获取中文摘要

1 Project Summary NIH is increasing its investment in large, multi-center brain MRI studies via projects such as the recently announced BRAIN initiative. The success of these studies depends on the quality of MRIs and the resulting image measurements, regardless of sample size. Even though quality control of MRIs and corresponding measurements could be outsourced, most neuroscience studies rely on in-house procedures that combine automatically generated scores with manually guided checks, such as visual inspection. Implementing these procedures typically requires combining several software systems. For example, the NIH NIAAA- and BD2K- funded Data Analysis Resource (DAR) of the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) uses XNAT to consolidate the structural, diffusion, and functional MRIs acquired across five sites, and has also developed their own custom software package to comply with study requirements for a multi-tier, quality control (QC) workflow. However, these custom, one-off tools lack support for the multi-site QC workflows that will come with the unified platform that MIQA represents: a design that supports collaboration and sharing, and strong cohesion between technologies. To improve the effectiveness of QC efforts specific to multi-center neuroimaging studies, we will develop a widely accessible and broadly compatible software platform that simplifies the creation of custom QC workflows in compliance with study requirements, provides core functionality for performing QC of medical images, and automatically generates documentation compliant with the FAIR principle, i.e., making scientific results findable, accessible, interoperable, and reusable. Specifically, our multi-site, web-based software platform for Medical Image Quality Assurance (MIQA) will enable efficient and accurate QC processing by leveraging open-source, state-of-the-art web interface technologies, such as a web-based dataset caching system and machine learning to aid in QC processes. Users will be able to configure workflows that not only reflect the specific requirements of medical imaging studies but also minimize the time spent on labor-intensive operations, such as visually reviewing scans. Issue tracking technology will enhance communication between geographically-distributed team members, as they can easily share image annotations and receive automated notifications of outstanding QC issues. The system will be easy to deploy as it will be able to interface with various imaging storage backends, such as local file systems and XNAT. While parts of this functionality have been developed elsewhere, MIQA is unique as it provides a unified, standard interface for efficient QC setup, maintenance, and review for projects analyzing multiple, independently managed data sources. The usefulness of this unique QC system will be demonstrated on increasing the efficiency of the diverse QC team of the multi-center NCANDA study.