Synthetic DCE-MRI Data

From QIBA Wiki
Jump to: navigation, search

Dr. Barboriak's Synthetic DCE-MRI Data - Access Instructions

The purpose of this effort is to produce synthetic DCE-MRI data appropriate for performing early stage verification of DCE-MRI analysis software. For our lab's first release of data in early July we have provided noise-free image data in DICOM format created using the extended Toft's model for a variety of vp, Ktrans and ve combinations.


Here are instructions for logging into a web page to get access to DCE-MRI test images. Apologies in advance -- it is a two step process, requiring human intervention at each step, so access is not immediate.


  • Step 1: Get a password for the website:

First, browse to the home page for my lab: http://dblab.duhs.duke.edu/ Second, click the "Login/Register" button at the bottom of the left hand column. Third, on the Login/Registration page, under User Login, click the "Click here to register for access to this site" link. Fourth, choose a password to go with your email address to use to log onto the site. You will be asked to respond to an email from Duke in order to confirm your login/password combination.

This is in part a manual process, so at this point wait for email from us telling you you've successfully registered (we'll likely get to you within 24 hours).


  • Step 2: Get access to the DCE-MRI Test Images

After you get email saying you have successfully registered for the lab website, login to our website with your email address/password. Click on the "Contact Us" link on the left side of the lab homepage. In the Comments field type in keyword "QIBA" and in the message text give us your name and institution/company, or something to identify you as a QIBA member. When you receive a confirmatory email, you will be able to download the DCE-MRI test data. A link will appear on the right side of this page that will allow the test data to be downloaded.


How to get to the right web page using your password: Once we give you access to the site, you will be able to use the email/pass combination to access the site directly at this address: http://dblab.duhs.duke.edu/modules/QIBAcontent/index.php?id=1


Alternatively, the page is also a link under the QIBA projects heading on the home page for my lab, on the right side.


What is on the website: A description of the test data. A .zip file of the test data in DICOM format.

QIBA Test Data Update - 2009.03.05

David Clunie, MBBS, has kindly provided some advice about the details of DICOM header formation for the simulated image data. A new version of the QIBA test data (QIBA_v3) has been put on the website.

Changes include:

  • Headers improved to pass Dr. Clunie's validator software.
  • Both DICOM and XML versions of the images provided. XML versions have the advantage of allowing the DICOM tag values to be edited using a text editor. New DICOM images can be easily generated using dcm4che's tool called "xml2dcm," available at: http://www.dcm4che.org/confluence/display/d2/dcm4che2+DICOM+Toolkit.
  • No changes to the images at this point.


The new images are available at: https://dblab.duhs.duke.edu/modules/QIBAcontent/index.php?id=1


  • Daniel Barboriak, MD (Dr. Barboriak)
    • Neuroradiology
    • Duke University Medical Center
    • barbo013[at]mc.duke.edu

Working Documents

Call Summaries