Abstract
Objectives To create and integrate a dedicated MSBase Imaging Repository (MSBIR) with the MSBase registry. To recruit MSBase sites to contribute to MSBIR. To facilitate quantitative analysis of brain MRI scans uploaded to MSBIR using an automated AI-based software platform and transmit these metrics into MSBase.
Methods Following stake-holder consultation, technical work commenced on the MSBIR build by Radiologics (Extensible Neuroimaging Archive Toolkit [XNAT] experts), Sydney Neuroimaging Analysis Centre (SNAC) & University of Sydney (USYD). Multiple sites were contacted regarding contribution to MSBIR. Automated, AI-based imaging pipelines measuring cross-sectional and longitudinal brain lesion and volume metrics from compatible clinically-acquired multiple sclerosis (MS) MRI scans were developed and refined by SNAC.
Results The customised MSBIR-XNAT production release was deployed on USYD hosted Amazon Web Services servers. SNAC developed products support the platform: (i)TORANATM, medical image gateway service that de-identifies images over secure/encrypted protocols, (ii)COEUSTM, advanced-search web portal, allows data retrieval for MSBase projects. Currently 7 sites have contributed 16235 MRI scans to MSBIR. Fully-automated quantitative analysis pipelines have been developed and implemented. Quantitative MRI brain lesion and volume metrics are available for compatible scans. All metrics are: (i)Stored in MSBIR/MSBase (de-identified data), (ii)Displayed in corresponding MSBase patient records.
Conclusions MSBIR-XNAT has been deployed and integrated with MSBase. Imaging data ingestion and further site/subject recruitment is ongoing. Compatible MRI brain scans entering MSBIR are automatically analysed by imaging pipelines and quantitative data stored in MSBIR and displayed in MSBase patient records. Clinical-imaging MS research collaborations utilising MSBIR are underway.
demonstrates a rare case of intramedullary spinal cord metastasis from melanoma with an acute thoracic cord syndrome.
