questions
58 Subscribers 
questions > sheared T1´s after DICOM to Nifti conversion
Showing 1-3 of 3 posts
Feb 2, 2017 03:02 PM | Alexandra Martinez
sheared T1´s after DICOM to Nifti conversion
Dear experts,
I am converting some DICOM files acquired through a Philips Achieva 3T MRI system, I have been converting these to NIFTI format on a High Computing Cluster using dcm2nii. However, we have recently identified a shear on the T1 NIFTI images. This shear is not present in the DICOM files, so it must be a conversion issue. I have tried the different dcm2nii versions, including the latest 2016 version and dcm2niix through microGL. We were told by a Philips specialist that dcm2nii/dcm2niix somehow applied a transformation/scaling value of 1 to the slice dimension (should be 0) and that this resulted in our slice dimension being 0.58 instead of 0.9 which would be the right slice dimension. This has resulted on the scaling factors on all 3 axes and the origin being changed. I was wondering if anyone has encountered this issue before and if there is a way of solving it. I have also realized that there has been a new release of dcm2niix in January but it seems that the compiled version isn't available just yet.
Any help would be greatly appreciated.
Thanks,
Alexandra
I am converting some DICOM files acquired through a Philips Achieva 3T MRI system, I have been converting these to NIFTI format on a High Computing Cluster using dcm2nii. However, we have recently identified a shear on the T1 NIFTI images. This shear is not present in the DICOM files, so it must be a conversion issue. I have tried the different dcm2nii versions, including the latest 2016 version and dcm2niix through microGL. We were told by a Philips specialist that dcm2nii/dcm2niix somehow applied a transformation/scaling value of 1 to the slice dimension (should be 0) and that this resulted in our slice dimension being 0.58 instead of 0.9 which would be the right slice dimension. This has resulted on the scaling factors on all 3 axes and the origin being changed. I was wondering if anyone has encountered this issue before and if there is a way of solving it. I have also realized that there has been a new release of dcm2niix in January but it seems that the compiled version isn't available just yet.
Any help would be greatly appreciated.
Thanks,
Alexandra
Feb 2, 2017 04:02 PM | Chris Rorden
RE: sheared T1´s after DICOM to Nifti conversion
I have never seen this, and do not have access to a Philips system.
Can you send me a sample dataset via dropbox to my private email?
Feb 21, 2017 01:02 PM | Chris Rorden
RE: sheared T1´s after DICOM to Nifti conversion
Thanks for the example files. It looks like my software is working
correctly. There are two comments:
1.) Your T2 scan has an anisotropic resolution, with 0.45mm in plane and 5mm between voxel centers. My software records this correctly, but different image viewer may decide to either scale images proportionally by interpolating in the slice direction (so the image does not appear spatially distorted) or to show the raw voxel data (which is better if you want to draw lesions or image masks). Some tools like MRIcron allow users to choose their preference for these images.
2.) The survey scan appears funny because it uses different slice orientations for the same series. These localizer scans are only useful during acquisition, as these scout scans allow the MRI technologist to determine the position and angulation of the subsequent high resolution scans. These localizers are not useful for subsequent analysis due to low resolution, partial sampling, missing slices, non-orthogonal slices, etc and so should be discarded.
1.) Your T2 scan has an anisotropic resolution, with 0.45mm in plane and 5mm between voxel centers. My software records this correctly, but different image viewer may decide to either scale images proportionally by interpolating in the slice direction (so the image does not appear spatially distorted) or to show the raw voxel data (which is better if you want to draw lesions or image masks). Some tools like MRIcron allow users to choose their preference for these images.
2.) The survey scan appears funny because it uses different slice orientations for the same series. These localizer scans are only useful during acquisition, as these scout scans allow the MRI technologist to determine the position and angulation of the subsequent high resolution scans. These localizers are not useful for subsequent analysis due to low resolution, partial sampling, missing slices, non-orthogonal slices, etc and so should be discarded.