Importing DICOM from a PACS via the DQR plugin

The Clara Train software operates on files that are in NIFTI format. Medical images are generated by acquisition modalities using the DICOM standard and are uploaded into XNAT through a variety of mechanisms that are described on this page: Image Session Upload Methods in XNAT.

XNAT has a DICOM Query Retrieve (DQR) plugin that was released in May of 2020 and facilitates retrieval and de-identification of images that are stored in a clinical or research PACS. Full documentation on that plugin is found here: DICOM Query-Retrieve Plugin

In brief, this plugin allows the user to

  • Search the PACS interactively to find imaging studies to retrieve, or
  • Create a CSV file with search criteria to identify individual imaging studies and additional columns to remap select identifiers when the images are imported into XNAT
  • Retrieve the images from the PACS into a selected project

Later sections will describe the process of converting DICOM images to NIFTI. The conversion process will rely on content found in the DICOM metadata. DICOM data is often acquired in separate series that are represented in XNAT as separate scans. The DICOM Series Description element is mapped to the XNAT Scan Description field.

When converting the data to NIFTI format, the tools described below are able to perform criteria matching on the Scan Description and other XNAT elements. The process is simplified if your data has regular or uniform text in the DICOM Series Description field. Imaging studies with multiple series have a further consideration. You might be studying one series from each study with common acquisition parameters such as an MR pulse sequence or CT data with a specific reconstruction sequence. You can choose to import only that data using the common XNAT tools or DQR plugin or import all of the data and select the appropriate scans later in the process. There will certainly be cases where that information is not known up front and you will import all of the images and filter the results later.

$label.name