Hope you are well. We have recently tried to use the generic protocol for MRI acquisition of spinal cord. As I checked the acquired images by radiologist, there are several and separate DWI acquisitions, including: DWI_b0, 2 DWI, DWI_ADC, DWI_TRACEW, DWI_FA, DWI_ColFA. The first issue is that how all these could create just one and unique DWI when we convert them to NIfTI from dicom? secondly, if we would like to only use DWI and DWI_b0 (which I supposed this should be right), how we can employ them for analysis? I have tried to merge DWI and DWI_b0 dicoms and then convert them all to NIfTI, which gives me two NIfTI files (dwi and dwi_b0), one bvec, one bval and two json files. However, when I try to analyze them using SCT, it returns warning and error as following:
WARNING: bvecs file is 3xn instead of nx3. Consider using sct_dmri_transpose_bvecs.
Transpose bvecs…
ERROR: no b=0 images detected. Maybe you are using non-null low bvals? in that case use flag -bvalmin. Exit program.
I greatly appreciate if you guide me in this regard, so I can execute the SCT commands for DWI properly.
Hope you are well. We have recently tried to use the generic protocol for MRI acquisition of spinal cord.
Great
As I checked the acquired images by radiologist, there are several and separate DWI acquisitions, including: DWI_b0, 2 DWI, DWI_ADC, DWI_TRACEW, DWI_FA, DWI_ColFA.
no, these are not different acquisitions but different reconstructions-- Siemens fits the DTI model and outputs the ADC, FA, etc. alongside the DWI data. Talk to your MR tech/radiologist, they should know about that. This is pretty standard.
The first issue is that how all these could create just one and unique DWI when we convert them to NIfTI from dicom?
see previous answer
secondly, if we would like to only use DWI and DWI_b0 (which I supposed this should be right), how we can employ them for analysis?
it depends which protocol you have used-- if you used the Trio/Verio protocol, then there is indeed an additional b=0 scans that you need to concatenate. If you used the skyra/prisma/vida there is no additional b=0 scan, so no need for concatenation. You can get inspiration from this code, which works for all vendors/models for dealing with additional b=0 scans.
I have tried to merge DWI and DWI_b0 dicoms and then convert them all to NIfTI, which gives me two NIfTI files (dwi and dwi_b0), one bvec, one bval and two json files. However, when I try to analyze them using SCT, it returns warning and error as following:
that’s not the correct way of doing it-- don’t merge dicom files. See my previous answer with link to code.
Hi @jcohenadad I am still working on it (I am updating some parts of my previous code and modify the DWI section from the code you suggested). Once I finished the analysis will inform you under this topic whether the issue solved or not.
