Time for DICOM to accommodate advances in MR technology

Even if you only occasionally follow the ups and downs of magnetic resonance technology, it is obvious that some major advances have occurred in MR in recent years. Given these improvements, it seems appropriate to reconsider the content of the MR information object defined by the DICOM committee in the early 1990s and to address new technologies using a more flexible interface.

Some users, for example, have an interest in the raw data that is used to generate images for reprocessing or in keeping the data for when new processing techniques become available. They like to export this information and process it at dedicated MR workstations. Other users would like to see the spectroscopy data as well because they can reveal information that is not necessarily obvious from the reconstructed images.

The variety of image processing and acquisition techniques has expanded as well. It is now possible to change parameters on the fly during acquisition and get information based on different physical characteristics. Initially the only dimensions available for acquired images were time and physical location; now the number of dimensions is almost infinite when all the various acquisition possibilities are taken into account.

The DICOM MR working group is focusing on an extension of the MR object definition to accommodate these and other developments. This is important from a user perspective for several reasons. First, it will eliminate the need for the private exchange of diagnostic information such as spectroscopic data by allowing this data to be properly stored and displayed on other vendors' workstations. (One caveat: You have to make sure the system supports the new DICOM SOP Class-this information is available in the DICOM conformance statement.)

The MR object definition extensions will also enable the exchange of spectroscopic and raw data, not just image data. In addition, image data will no longer be exchanged as individual images but as multiframe images. It is easy to package all pixel data of a full image acquisition, such as a cine loop, into a single folder. The trick is how to describe the differences between the individual frames of the multiframe object. The DICOM committee has overcome this problem by developing a convenient description of all data dimensions.

Final approval of the proposed MR DICOM extension is expected this summer. In the meantime, the proposed changes are available for public comment. Interested parties can get a copy of the proposal from the NEMA Web site (www.nema.org).

People sometimes ask me when the DICOM standard will finally be done. The answer is that as long as the modalities keep evolving, there will always be a need to accommodate these changes. Vendors will always have to add new features and changes to their interfaces, and users will continue to have to upgrade their equipment. Good news or bad news? Well, progress requires a certain effort.