Documentation/Nightly/FAQ/DICOM

For the latest Slicer documentation, visit the read-the-docs.

DICOM

What is DICOM anyway?

Digital Imaging and Communications in Medicine (DICOM) is the standard for the communication and management of medical imaging information and related data (see Wikipedia article here). Among other things, DICOM defines format of objects for communicating images and image-related data. In most cases, imaging equipment (CT and MR scanners) used in the hospitals will generate images saved as DICOM objects.

DICOM organizes data following the hierarchy of

• Patient ... can have 1 or more
  • Study (single imaging exam encounter) ... can have 1 or more
    • Series (single image acquisition, most often corresponding to a single image volume) ... can have 1 or more
      • Instance (most often, each Series will contain multiple Instances, with each Instance corresponding to a single slice of the image)

As a result of imaging exam, imaging equipment generates DICOM files, where each file corresponds to one Instance, and is tagged with the information that allows to determine the Series, Study and Patient information to put it into the proper location in the hierarchy.

There is a variety of DICOM objects defined by the standard. Most common object types are those that store the image volumes produced by the CT and MR scanners. Those objects most often will have multiple files (instances) for each series. Image processing tasks most often are concerned with analyzing the image volume, which most often corresponds to a single Series. The first step working with such data is to load that volume into 3D Slicer.

How do I know if the files I have are stored using DICOM format? How do I get started?

DICOM files do not need to have a specific file extension, and it may not be straightforward to answer this question easily. However, if you have a dataset produced by a clinical scanner, it is most likely in the DICOM format. If you suspect your data might be in DICOM format, it might be easiest to try to load it as DICOM:

1) drag and drop the directory with your data into Slicer window. You will get a prompt "Select a reader to use for your data? Load directory into DICOM database." Accept that selection. You will see a progress update as the content of that directory is being indexed. If the directory contained DICOM data, and import succeeded, at the completion you will see the message of how many Patient/Study/Series/Instance items were successfully imported.

2) Once import is completed, you will see the window of the DICOM Browser listing all Patients/Studies/Series currently indexed. You can next select individual items from the DICOM Browser window and load them.

3) Once you load the data into Slicer using DICOM Browser, you can switch to the "Data" module to examine the content that was imported.

When I click on "Load selection to slicer" I get an error message "Could not load ... as a scalar volume"

A common cause of loading failure is corruption of the DICOM files by incorrect anonymization. Patient name, patient ID, and series instance UID fields should not be empty or missing (the anonymizer should replace them by other valid strings). Try to load the original, non-anonymized sequence and/or change your anonymization procedure.

If none of the above helps then check the Slicer error logs and report the error on the Slicer forum. If you share the data (e.g., upload it to Dropbox and add the link to the error report) then Slicer developers can reproduce and fix the problem faster.

I try to import a directory of DICOM files, but nothing shows up in the browser

DICOM is a complex way to represent data, and often scanners and other software will generate 'non-standard' files that claim to be DICOM but really aren't compliant with the specification. In addition, the specification itself has many variations and special formats that Slicer is not able to understand. Slicer is used most often with CT and MR DICOM objects, so these will typically work.

If you have trouble importing DICOM data here are some steps to try:

• Make sure you are following the DICOM module documentation.
• We are constantly improving the application (new preview version is released every day), so there is a chance that the problem you encountered is addressed in a recent version. Try loading the data using the latest stable and the latest nightly versions of Slicer.
• Make sure the Slicer temporary folder is writeable. Temporary folder can be selected in menu: Edit / Application Settings / Modules / Temporary directory.
• Try moving the data and the database directory to a path that includes only US English characters (ASCII) to avoid possible parsing errors. No special, international characters are allowed.
• To confirm that your installation of Sicer is reading data correctly, try loading other data, such as this anonymized sample DICOM series (CT scan)
• Try import using different DICOM readers: in Application settings / DICOM / DICOMScalarVolumePlugin / DICOM reader approach: switch from DCMTK to GDCM (or GDCM to DCMTK), restart Slicer, and attempt to load the data set again.
• See if the SlicerDcm2nii extension will convert your images. You can install this module using the Extension manager. Once installed you will be able to use the Dcm2niixGUI module from slicer.
• Try the DICOM Patcher module.
• Review the Error Log for information.
• Try loading the data by selecting one of the files in the Add Data Dialog. Note: be sure to turn on Show Options and then turn off the Single File option in order to load the selected series as a volume
• If you are still unable to load the data, you may need to find a utility that converts the data into something Slicer can read. Sometimes tools like FreeSurfer, FSL or MRIcron can understand special formats that Slicer does not handle natively. These systems typically export NIfTI files that slicer can read.
• For archival studies, are you sure that your data is in DICOM format, or is it possible the data is stored in one of the proprietary MR or CT formats that predated DICOM? If the latter, you may want to try the dcm2nii tool distributed with MRIcron up until 2016. More recent versions of MRIcorn include dcm2niix, which is better for modern DICOM images. However, the legacy dcm2nii includes support for proprietary formats from GE, Philips, Siemens and Elscint.
• If you are sure that the DICOM files do not contain patient confidential information, you may post a sample dataset on a web site and ask for help from Slicer forum. Please be careful not to accidentally reveal private health information. If you want to remove identifiers from the DICOM files you may want to look at DicomCleaner, gdcmanon or the RSNA Clinical Trial Processor software.

Something is displayed, but it is not what I expected

I would expect to see a different image

When you load a study from DICOM, it may contain several data sets and by default Slicer may not show the data set that you are most interested in. Go to Data module / Subject hierarchy section and click the "eye" icons to show/hide loaded data sets. You may need to click on the small rectangle icon ("Adjust the slice viewer's field of view...") on the left side of the slice selection slider after you show a volume.

If none of the data sets seems to be correct then follow the steps described in section "I try to import a directory of DICOM files, but nothing shows up in the browser".

Image is stretched or compressed along one axis

Some non-clinical (industrial or pre-clinical) imaging systems do not generate valid DICOM data sets. For example, they may incorrectly assume that slice thickness tag defines image geometry, while according to DICOM standard, image slice position must be used for determining image geometry. DICOM Patcher module can fix some of these images: remove the images from Slicer's DICOM database, process the image files with DICOM Patcher module, and re-import the processed file into Slicer's DICOM database. If image is still distorted, go to Volumes module, open Volume information section, and adjust Image spacing values.

Scanners may create image volumes with varying image slice spacing. Slicer can represent such images in the scene by apply a non-linear transform. To enable this feature, go to menu: Edit / Application settings / DICOM and set Acquisition geometry regularization to apply regularization transform. Slice view, segmentation, and many other features work directly on non-linearly transformed volumes. For some other features, such as volume rendering, you need to harden the transform on the volume: go to Data module, in the row of the volume node, double-click on the transform column, and choose Harden transform.

Note that if Slicer displays a warning about non-uniform slice spacing then it may be due to missing or corrupted DICOM files. There is no reliable mechanism to distinguish between slices that are missing because they had not been acquired (for example, to reduce patient dose) or they were acquired but later they were lost.