sbonaretti/pyKNEEr

Calculation of T2 maps

Opened this issue · 3 comments

FroylanZR commented

Hi Serena,

Im using my own set of images and am in the relaxometry step. Right now i only had one data set of the studies, so im not sure if the 3D map i got is accurate. I used the following files in my image_list_relaxometry_EP.txt file:

./preprocessed
./segmented
i1 01_S1T2_orig.mha
i2 01_S1T2_prep.mha
cm 01_S1T2_prep_fc.mha

sbonaretti commented

Hi!

I am finally back to full internet connection - thanks for your patience!

To calculate relaxometry maps (T2 or T1rho) you should use only *_orig.mha images because they contain the original image intensities (relaxometry maps are calculated from image intensities).
*_prep.mha images contain standardized intensities to favor segmentation, thus they should not be used to calculate relaxometry maps.
In addition, *_orig.mha and *_prep.mha images have the same mask (*_prep.fc.mha), because they went through the same spatial preprocessing, i.e. they have the same image orientation (RAI), origin (0,0,0) and laterality (left). So it correct to use *_prep_fc.mha also for *_orig.mha images

To calculate the relaxometry maps with pyKNEEr you need specific image sequences. To calculate:

  • T2, you need a DESS acquisition, which contains 2 images acquired in the same acquisition
  • T1rho, you need a cubeQuant acquisition, which contains 4 images acquired in the same acquisition

What kind of images are you using?

FroylanZR commented

Hi Serena,

We am currently using T2 images and trying to get optimal T1rho images to work with as well.

We are starting a new Knee project and we have a Siemens Vida in house with these sequences. We do not have T1rho.

• T2, you need a DESS acquisition, which contains 2 images acquired in the same acquisition
o We have Siemens Knee Cartilage: 3D DESS 160 slices; 0.60 mm [t2_de3d_we_sag_iso]
• T1rho, you need a cubeQuant acquisition, which contains 4 images acquired in the same acquisition
o Can we use this? T2*-weighted 3D Multi-Echo Data Image Combination [t2_me3d_we_sag_iso]

any feedback would be appreciated! thanks!

sbonaretti commented

Hi,

I am not familiar about your image acquisition protocols, so I unfortunately cannot give you much of a feedback. But here are some information related to what I used, so maybe you can find similarities and differences with your protocols:

  • pyKNEEr's paper: at page 9, paragraph "Cartilage relaxometry" you can see how I calculated the maps
  • Linear fitting paper: describing the exponential fitting and the acquisition protocols of the used images
  • EPG paper: describing calculation of T2 maps from DESS images using Extended Phase Graph (EPG) modeling and the acquisition protocols of the used images

Hope this can be of help!