Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Joining together data from two different MRI protocols

  1. Oct 19, 2015 #1


    User Avatar
    Gold Member

    I should start by saying that, as a novice data analyst, I have very little experience with MRI physics, but I believe I understand the absolute fundamentals. Also, this post mostly concerns data analysis issues so might be better suited for some other signal processing forum, but I'm hoping someone knowledgeable about MRIs might be able to at least point me in the right direction.

    Below is as much as I have been able to understand about the data, assuming it's correct.

    I have been given two data sets, encompassing a total three hour contrast enhanced MRI scan. In the first ten minutes, a large quantity of 'DCE' scans are taken, with contrast injection taking place sometime after the first ten or eleven DCE samples. These ten or eleven samples are pre-injection, and thus used as a baseline.

    This produces some signal intensity for each voxel, which we can convert to a T1 weighted volume using the following equation:

    [tex] S = \frac{S_0 sin(\alpha) (1 - E_1) e^{-TE/T2}}{[1-E_1 cos(\alpha)]}[/tex]

    This equation is taken from an older paper: Contrast in Rapid MR Imaging: T1- and T2-Weighted Imaging by Richard B. Buxton, et al, and was given to me by the MRI scientists responsible for designing the protocol. They have verified that this is an equation I need to use.

    From what I've gathered so far by talking to them:
    • since the TE value is very near 0, the equation simplifies.
    • [itex]S_0[/itex] for a given voxel is the average value of the DCE values there, pre-injection.
    So, the remaining data set is not DCE, but standard 'T1-maps' (whatever the difference is between DCE and T1-maps I am still not entirely sure..., apparently each T1-map value is actually a value derived from 5 different scans with different flip angles, but I was not told how this was done) and supposedly the above equation can help me transform the DCE data into T1-maps, thereby allowing me to join the two data sets.

    A T1-map has been taken pre-injection ([itex]T_{1,0}[/itex])also, to aid me in adjusting the DCE-derived T1 values so the two data sets agree on the same baseline, but I have been given no instructions on how to do this, and the MRI scientists appear uninterested in helping me understand how to process this data.

    Once I have the entire data set in agreement on units and normalized to the correct baseline, I can calculate the concentration of the contrast agent in each voxel, which is what I'm really trying to get at here.

    I have tried two different linear transformations to match the baselines but neither have been correct, and I am not sure if this baseline normalization is even supposed to be a linear transformation or not. After everything is said and done, the curve is supposed to appear smoothly decreasing, but there is a definite discontinuity between the DCE-derived T1 values and the T1-maps.

    Perhaps there is something I am completely missing here, I welcome any suggestions at this point.
  2. jcsd
  3. Oct 24, 2015 #2
    Thanks for the post! This is an automated courtesy bump. Sorry you aren't generating responses at the moment. Do you have any further information, come to any new conclusions or is it possible to reword the post?
  4. Nov 23, 2015 #3


    User Avatar
    Gold Member

    This issue has been mostly resolved now, I will hopefully update this post with more information when I have the time.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook