Calculation of CBF and ATT map from multi Delay ASL

39 views
Skip to first unread message

s4s...@gmail.com

unread,
May 17, 2024, 1:45:20 AM5/17/24
to ASLtbx discussion board
Sir,
I have multiPLD Perfusion images with different PLD values(1525ms, 2025ms, 2525ms and 3025ms), PD image and Structural image obtained from GE scanner. How can i use ASLtbx  script  to calculate CBF and ATT map from Perfusion images instead of L-C pairs.

red...@gmail.com

unread,
May 24, 2024, 6:59:40 AM5/24/24
to asltbx-disc...@googlegroups.com, Yiran Li

You can have two options: one is the weighted delay based approach, the second is the kinetic model fitting based approach.  The first approach depends on the assumption of the relationship between the weighted sum and the real transit time which may not be accurate. But as long as you are using the same scripts and the same model, the results are comparable across different subjects.  The second approach is more accurate but the fitting can be unstable.  You can contact Yiran here for the fitting based scripts.  For the weighted delay approach, you can find the code in this discussion board.

 

Ze

--
You received this message because you are subscribed to the Google Groups "ASLtbx discussion board" group.
To unsubscribe from this group and stop receiving emails from it, send an email to asltbx-discussion...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/asltbx-discussion-board/f7b5d1f5-d6f8-4645-9067-7595a8faf99an%40googlegroups.com.

SHYNA A

unread,
May 26, 2024, 12:23:16 PM5/26/24
to ASLtbx discussion board

Thank you sir for your suggestion. I am a beginner in Multi PLD approach. Based on the suggestions from you, I tried to find the ATT map based on the data obtained from GE scanner . I have used perfusion images with PLD=1525, 2025, 2525 and 3025. I wrote simple script t as shown below to understand the output of each step. Kindly see the attached  ATT map obtained for this script. Is it correct? I coudlnot find the problem   with the following script.  I am also attaching the input data that I have used. Please  give a suggestion to get the correct ATT map.


   perfmaps=0;
    sumperf=0;
    cbfmap=0
    v1=spm_vol('E:\multi_PLD\data\sub11\meanPERF_1525.nii');
    dat=spm_read_vols(v1);
    sumperf=sumperf+dat;
    perfmaps=perfmaps+ 1.525*dat;
    v2=spm_vol('E:\multi_PLD\data\sub11\meanPERF_2025.nii');
    dat=spm_read_vols(v2);
    sumperf=sumperf+dat;
    perfmaps=perfmaps+ 2.025*dat;
    v3=spm_vol('E:\multi_PLD\data\sub11\meanPERF_2525.nii');
    dat=spm_read_vols(v3);
    sumperf=sumperf+dat;
    perfmaps=perfmaps+ 2.525*dat;
    v4=spm_vol('E:\multi_PLD\data\sub11\meanPERF_3025.nii');
    dat=spm_read_vols(v4);
    sumperf=sumperf+dat;
    perfmaps=perfmaps+ 3.025*dat;
    vo=v1;
    vo.fname=fullfile(PAR.root,PAR.subjects{1}, ['perfmaps_' PAR.subjects{1} '.nii']);
    vo=spm_write_vol(vo,perfmaps); 
    vo.fname=fullfile(PAR.root,PAR.subjects{1}, ['sumperf1_' PAR.subjects{1} '.nii']);
    vo=spm_write_vol(vo,sumperf); 
    wdm=perfmaps./sumperf;
    att=1.761*wdm-1.532;
    vo.fname=fullfile(PAR.root,PAR.subjects{1}, ['wdm_' PAR.subjects{1} '.nii']);
    vo=spm_write_vol(vo,att); 
    vo.fname=fullfile(PAR.root,PAR.subjects{1}, ['attmap_' PAR.subjects{1} '.nii']);
    vo=spm_write_vol(vo,att); 

attmap.PNG

ATT map





Perfmapsperfmaps.PNG

red...@gmail.com

unread,
Jun 1, 2024, 9:06:35 AM6/1/24
to asltbx-disc...@googlegroups.com

The process is right. I didn’t see any problem in your scripts. Make sure you used the perfusion difference maps rather than CBF images in the calculation.  The equation: att=1.761*wdm-1.532 was based on simulation and then data fitting, which might be not accurate. But you should still be able to get a relative ATT map. 

Could you also check the image scale of your att map? Could you display it using a display window of 0-2.5 sec?

 

ATT map

 

 

 

 

 

Perfmaps

 

 

 

On Friday, May 24, 2024 at 4:29:40PM UTC+5:30 red...@gmail.com wrote:

You can have two options: one is the weighted delay based approach, the second is the kinetic model fitting based approach.  The first approach depends on the assumption of the relationship between the weighted sum and the real transit time which may not be accurate. But as long as you are using the same scripts and the same model, the results are comparable across different subjects.  The second approach is more accurate but the fitting can be unstable.  You can contact Yiran here for the fitting based scripts.  For the weighted delay approach, you can find the code in this discussion board.

 

Ze

 

From: asltbx-disc...@googlegroups.com <asltbx-disc...@googlegroups.com> On Behalf Of s4s...@gmail.com
Sent: Friday, May 17, 2024 1:45 AM
To: ASLtbx discussion board <asltbx-disc...@googlegroups.com>
Subject: Calculation of CBF and ATT map from multi Delay ASL

 

Sir,

I have multiPLD Perfusion images with different PLD values(1525ms, 2025ms, 2525ms and 3025ms), PD image and Structural image obtained from GE scanner. How can i use ASLtbx  script  to calculate CBF and ATT map from Perfusion images instead of L-C pairs.

--
You received this message because you are subscribed to the Google Groups "ASLtbx discussion board" group.
To unsubscribe from this group and stop receiving emails from it, send an email to asltbx-discussion...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/asltbx-discussion-board/f7b5d1f5-d6f8-4645-9067-7595a8faf99an%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "ASLtbx discussion board" group.
To unsubscribe from this group and stop receiving emails from it, send an email to asltbx-discussion...@googlegroups.com.

image001.png
image002.png
Reply all
Reply to author
Forward
0 new messages