Purpose or Case Report: Many diseases in pediatrics are relatively uncommon and imaging of these diseases is difficult to study due to small patient numbers at a single site. This limits the ability of single institutions to adequately power a study. Thus, there is a pressing need for a multi-site structure to combine data for rare diseases to appropriately power outcome studies. Our goal is to create an infrastructure to support pooling of imaging and clinical data across institutions to facilitate multi-institutional studies.
Methods & Materials: For multi-site studies, a centralized IRB structure is necessary to efficiently perform the study. The site establishing the centralized IRB will be considered the ‘primary’ site. The primary site is also responsible for IRB regulatory, data management, including imaging, and facilitation of project overall. The centralized IRB will establish reliance agreements with participating sites, enabling the sites to utilize the central IRB as the IRB of record. Once IRB approval is obtained, a centralized REDCap, or data entry system, at the primary site will grant access to participating institutions for data entry. Centralized data collection allows all data to be housed in one location, allowing for quicker analysis. For studies requiring imaging transfer, a standard operating procedure (SOP) for image de-identification, naming convention, and image transfer to centralized PACS system should be followed.
Results: Our site has established three multi-site clinical imaging trials to date. First, we have an 11-site imaging repository for pleuropulmonary blastoma through the Midwest surgical consortium. All images have been uploaded to our site and pediatric radiologists at three different institutions are viewing de-identified images and inputting data in REDCap. Second, we have a three-site retrospective study evaluating the use of grayscale and Doppler imaging in veno-occlusive disease. Each site has access to centralized REDCap and data entry is performed on local subjects. Third, we have a two-site retrospective and prospective fetal MRI registry. Data collection and image upload to centralized PACS is being used to combine data between institutions.
Conclusions: We have successfully created infrastructure to support multi-institution clinical imaging trials. We hope that SPR members can use this resource for future studies on rare conditions where we can pool data to see how imaging affects patient outcomes.