Back

 Back to Success Stories Mainpage
Back Back to What's New
Time-Resolved MR Angiography: Artery-Vein Separation using Correlation Analysis

Michael Bock, Deutsches Krebsforschungszentrum (dkfz), Heidelberg, Germany

With modern magnetic resonance (MR) scanners time-resolved 3D data sets of the blood vessels (MR angiographies) can be acquired in a single breath hold period. The series of 3D data sets shows the transit of contrast media (e.g. Gd-DTPA) through the vasculature and, thus, contains more information than static angiographies alone. In the lung, this dynamic information can e.g. be utilised to separate arterial from venous blood vessels. Since contrast media transit times through the lung can be very short, the individual 3D data sets often show signals from both arterial and venous vessels.

Analyze4D.jpg (33548 Byte)

MR angiography post-processing tool ANALYSE_4D

At the Deutsches Krebsforschungszentrum (German Cancer Research Center, DKFZ) the research group of Michael Bock has implemented an IDL tool, ANALYZE_4D, to post-process 4D MR angiography data sets. With this tool pure arterial and venous MR angiographies are computed from the many hundred original image files. To separate arteries from veins, a modified correlation analysis approach is used. After selection of a reference arterial or venous vessel the correlation analysis emphasizes only those vessel structures that show a similar enhancement pattern.

Artery-Vein Separation

Result of an artery-vein separation: lung arteries (blue) and veins (red) were separated by their distinctive contrast enhancement patterns.

For this project, the researchers at the DKFZ have chosen IDL as a development platform for various reasons:

  • The integrated DICOM support facilitates vendor-independent handling of the MR images.
  • The IDL virtual machine allows to disseminate the program to the clinical end users at the DKFZ with no additional cost for licenses.
  • The program can be used on a variety of host computers due to IDL's platform independence.
  • The graphics capabilities as well as the programmer-friendly GUI capabilites make IDL an excellent environment to develop both research-oriented and user-friendly applications.