questions > Warning about ADC Philips DATA
Showing 1-4 of 4 posts
Jul 22, 2016 09:07 PM | Alessandro Crippa
Warning about ADC Philips DATA
Dear colleagues,
I need to convert DTI sequence from a Philips scanner for analysis with FSL (basically, I need to create a nii.gz file and bval/bvec files).
I have converted our sequence using the dcm2nii GUI with the option "Compressed FSL (4D NIFTI nii)", and output files generated were the expected nii.gz and bval/bvec files.
During the conversion process I had this message:"Warning: final volume is computed ADC and will not be converted (as it would disrupt processing). You can re-create a better ADC image after eddy current correction".
However, the last volume in our sequence is the b0 (volume#34), not the ADC (#33). And so the nii.gz file generated does not include the b0 (but the ADC) (33 volumes in total).
After reading previous posts, I have tried to use the latest dcm2niix/MRIcroGL, but I could not call the GUI or open the conversion menu from MRIcroGL since I had this message: "libgcc_s_dw2-1.dll is missing".
How could I handle this point? What should I do to prevent the dcm2nii to cut away the b0 instead of the actual ADC?
Thank you in advance for any suggestions,
Best
Alessandro Crippa
I need to convert DTI sequence from a Philips scanner for analysis with FSL (basically, I need to create a nii.gz file and bval/bvec files).
I have converted our sequence using the dcm2nii GUI with the option "Compressed FSL (4D NIFTI nii)", and output files generated were the expected nii.gz and bval/bvec files.
During the conversion process I had this message:"Warning: final volume is computed ADC and will not be converted (as it would disrupt processing). You can re-create a better ADC image after eddy current correction".
However, the last volume in our sequence is the b0 (volume#34), not the ADC (#33). And so the nii.gz file generated does not include the b0 (but the ADC) (33 volumes in total).
After reading previous posts, I have tried to use the latest dcm2niix/MRIcroGL, but I could not call the GUI or open the conversion menu from MRIcroGL since I had this message: "libgcc_s_dw2-1.dll is missing".
How could I handle this point? What should I do to prevent the dcm2nii to cut away the b0 instead of the actual ADC?
Thank you in advance for any suggestions,
Best
Alessandro Crippa
Jul 23, 2016 02:07 PM | Chris Rorden
RE: Warning about ADC Philips DATA
1.) Can you send me an email with a link to your image. I do not
have access to Philips scanners, so I rely on users to validate the
software. A sample that demonstrates the problem would be a great
help. A sample that you are able to share publicly would be even
better as it would allow the developers of other converters
(dicm2nii, MRIconvert) to up their game.
2.) Are you converting a DICOM dataset or is this a Philips proprietawry PAR/REC. I am happy to tune my softwawre for either case, but strictly speaking if it is a DICOM dataset you should really contact Philips to update their software: including derived images (ADC, TRACE, FA) as the same series as the raw data is a clear violation of the DICOM standard.
3.) I will try to add the DLLs to the next release of MRIcroGL/dcm2niix. In the meantime you can find them with Google. However, since you want to use FSL which only runs on OSX or Linux (does not run on Windows) can I suggest that you use MRIcroGL/dcm2niix on OSX/Linux? I am a single individual maintaining a lot of tools: I have moved my development efforts from MRIcron/dcm2nii to MRIcroGL/dcm2niix. While my old tools are robust and mature, they can not foresee the numerous changes that are implemented by scanner vendors. Unfortunately, DICOM is a very complex format that is being continually re-interpreted by the manufacturers of imaging devices.
2.) Are you converting a DICOM dataset or is this a Philips proprietawry PAR/REC. I am happy to tune my softwawre for either case, but strictly speaking if it is a DICOM dataset you should really contact Philips to update their software: including derived images (ADC, TRACE, FA) as the same series as the raw data is a clear violation of the DICOM standard.
3.) I will try to add the DLLs to the next release of MRIcroGL/dcm2niix. In the meantime you can find them with Google. However, since you want to use FSL which only runs on OSX or Linux (does not run on Windows) can I suggest that you use MRIcroGL/dcm2niix on OSX/Linux? I am a single individual maintaining a lot of tools: I have moved my development efforts from MRIcron/dcm2nii to MRIcroGL/dcm2niix. While my old tools are robust and mature, they can not foresee the numerous changes that are implemented by scanner vendors. Unfortunately, DICOM is a very complex format that is being continually re-interpreted by the manufacturers of imaging devices.
Jul 28, 2016 03:07 PM | Alessandro Crippa
RE: Warning about ADC Philips DATA
Dear Chris,
thanks for your early reply.
I was converting a Philips PAR/REC file. I realized just yesterday that the dcm2nii software is working properly with our data, cutting away not the entire last volume (#33 in our philips data), but the last part of it (the ADC/trace, with b=700 but no gradients). So basically, the software was doing OK.
Thank you again for this software, it's a great product, powerful but ease to use even for beginners like me.
Best,
Alessandro
Originally posted by Chris Rorden:
thanks for your early reply.
I was converting a Philips PAR/REC file. I realized just yesterday that the dcm2nii software is working properly with our data, cutting away not the entire last volume (#33 in our philips data), but the last part of it (the ADC/trace, with b=700 but no gradients). So basically, the software was doing OK.
Thank you again for this software, it's a great product, powerful but ease to use even for beginners like me.
Best,
Alessandro
Originally posted by Chris Rorden:
1.) Can you send me an email with a link to your
image. I do not have access to Philips scanners, so I rely on users
to validate the software. A sample that demonstrates the problem
would be a great help. A sample that you are able to share publicly
would be even better as it would allow the developers of other
converters (dicm2nii, MRIconvert) to up their game.
2.) Are you converting a DICOM dataset or is this a Philips proprietawry PAR/REC. I am happy to tune my softwawre for either case, but strictly speaking if it is a DICOM dataset you should really contact Philips to update their software: including derived images (ADC, TRACE, FA) as the same series as the raw data is a clear violation of the DICOM standard.
3.) I will try to add the DLLs to the next release of MRIcroGL/dcm2niix. In the meantime you can find them with Google. However, since you want to use FSL which only runs on OSX or Linux (does not run on Windows) can I suggest that you use MRIcroGL/dcm2niix on OSX/Linux? I am a single individual maintaining a lot of tools: I have moved my development efforts from MRIcron/dcm2nii to MRIcroGL/dcm2niix. While my old tools are robust and mature, they can not foresee the numerous changes that are implemented by scanner vendors. Unfortunately, DICOM is a very complex format that is being continually re-interpreted by the manufacturers of imaging devices.
2.) Are you converting a DICOM dataset or is this a Philips proprietawry PAR/REC. I am happy to tune my softwawre for either case, but strictly speaking if it is a DICOM dataset you should really contact Philips to update their software: including derived images (ADC, TRACE, FA) as the same series as the raw data is a clear violation of the DICOM standard.
3.) I will try to add the DLLs to the next release of MRIcroGL/dcm2niix. In the meantime you can find them with Google. However, since you want to use FSL which only runs on OSX or Linux (does not run on Windows) can I suggest that you use MRIcroGL/dcm2niix on OSX/Linux? I am a single individual maintaining a lot of tools: I have moved my development efforts from MRIcron/dcm2nii to MRIcroGL/dcm2niix. While my old tools are robust and mature, they can not foresee the numerous changes that are implemented by scanner vendors. Unfortunately, DICOM is a very complex format that is being continually re-interpreted by the manufacturers of imaging devices.
Jul 28, 2016 03:07 PM | Chris Rorden
RE: Warning about ADC Philips DATA
Yes, this is the proper outcome. FSL and other tools need the raw
data, not the derived images (trace, ADC, MD, FA). Again, if these
are DICOM sequences, it is a clear violation of Philips to include
these derived images as the same series as the raw data. I hope
they fix this, but it is nice to know my software is doing the
right thing.