File(s) not publicly available
Reason: The weblink only directs people to the mechanism for requesting data access and not directly to the data itself.
A repository-integrated quantitative magnetic resonance imaging data analysis platform for enabling multi-centre clinical studies: application and measurable impact in an oxygen-enhanced MRI validation study of head and neck cancer
Variation in procedures for the transfer, analysis and quality control of data between sites in multi-centre imaging studies hinders reproducibility and makes translation of novel biomarkers to clinical practice more challenging. This work aimed to demonstrate a streamlined and harmonised workflow for data analysis and quality control that could be applied to magnetic resonance imaging data in a multi-centre clinical trial. A data processing pipeline reporting quantitative magnetic resonance imaging biomarkers, and the integration of these processes within the XNAT imaging repository using a software container platform was demonstrated in an exemplar patient with head-and-neck cancer imaged using a multi-parametric MRI scan. The developed container is highly customisable and facilitates a host of pre-processing techniques such as sorting, scaling and conversion of DICOM images, as well as image and region of interest registration. Data analysis can be performed to derive common quantitative imaging biomarkers, such as T1 relaxation time, apparent diffusion coefficient and various parameters obtained with intravoxel incoherent motion mapping, dynamic contrast-enhanced MRI, and more bespoke analyses such as oxygen-enhanced hypoxia mapping. The work presented here achieves this by utilising the Madym C++ quantitative MRI toolkit as its base. We provide an overview of data analysis container design and the interface within the image repository, along with a demonstration of user configuration options and example output. The main limitations of this study are that containerisation and integration of the software was performed using a single combination of container engine and imaging repository, and the workflow is yet to be tested in a multi-centre trial. Improvements to the repeatability of data acquisition and analysis from using a standardised, locked-down and traceable approach in multi-centre clinical studies, supports the translation of both novel and more established quantitative imaging biomarkers from research to clinical practice.
Access to data used in this study can be requested by following this link: https://qbi-xnat.manchester.ac.uk and clicking 'Register' or by emailing qbilab@manchester.ac.uk.