For longitudinal studies automatic voxel positioning seems to be interesting as this can improve reducibility significantly. Does anyone has (good/bad) practical experience with these techniques?
Some recent papers:
Bishop et al., 2021: Real-Time Semi-Automated and Automated Voxel Placement for Repeated Acquisition Magnetic Resonance Spectroscopy
Reid et al., 2021: https://doi.org/10.1002/nbm.4631
I’d also be interested in hearing anyone’s experience using these types of techniques. I know there was the AutoVOI work (AutoVOI: Real-time automatic prescription of volume-of-interest for single voxel spectroscopy - PMC) a few years ago, but I haven’t heard much since.
Haven’t tried AutoVOI, but it looks to be available : AutoVOI and still in use : https://onlinelibrary.wiley.com/doi/10.1002/mrm.29164.
Hello! I’ve had a good experience with AutoAlign on Siemens, which we used in the NMR Biomed paper. On the 7T MAGNETOM, the AutoAlign scout is 3 minutes, so it’s a very quick and easy way to get automatic repositioning for longitudinal studies, and I definitely recommend it.
I haven’t systematically tested AutoAlign for voxel placement across different subjects, but in my current cross-sectional study with a voxel in dorsolateral prefrontal cortex, my anecdotal experience is that AutoAlign gets the voxel close to where I want it, but I still need to make adjustments. I would be interested in trying out AutoVOI and other methods (e.g., https://doi.org/10.17756/jnpn.2018-020) for better reproducibility between subjects.
Yes, AutoAlign on Siemens scanner works pretty good to have the same MRS voxel location for longitudinal study. Only thing is that for the first scan, you will need to manually place the VOI or you could use AutoVOI, then save this protocol for the next scan for this particular subject.
Hi all,
This is a long shot, but we are getting a couple of longitudinal spectroscopy studies up and running, and we were hoping to use AutoAlign with the Siemens product svs sequences for reproducible voxel placement across sessions. We recently upgraded to XA30, though, and it seems that the svs sequences no longer include an AutoAlign option. We’re wondering if anyone else has run into this and/or can think of any workarounds we might try. Thank you!
Hi all,
Has anyone found a way to make AutoAlign work on XA30?
We were long time users on the VE11C and our long-delayed “upgrade” to XA30 has broken this for us as well.
We only need this for automated placement within participants in longitudinal studies. Manual placement at the first scan is no problem.
We ended up deciding it just wasn’t possible on XA30, so I’ve been working on a little python package that runs on a separate computer and solves just this problem. Using the T1 and spectroscopy DICOM(s) from the first scan session, it takes in the new T1 during a subsequent scan and calculates the voxel prescription that needs to be entered on the console to exactly match the voxel placement from the first scan session.
Our current documentation is available here. I’d be happy to consult if you’d be interested in giving it a try!
This sounds very useful! Is the github repo public? I get a 404 when I try and click through.
Did you manage to unambiguously work out the transform from NIfTI/DICOM back to the scanner prescription? It seems like it used some very unclear combination of internal/external Euler angles.
Hi @wclarke ! I’m waiting on confirmation from my university that I’m handling attribution/licensing correctly and then I’ll be able to make the repo public. I anticipate that’ll be settled within the next week. It would be great to get more sets of eyes on it!
At least in our hands, things seem to be working just fine; it’s possible that there are edge cases somewhere that lead to ambiguity in the translation, but we haven’t come upon them yet. If you have an intuition for tests that might expose those sorts of issues, please let me know!
I’ll be in touch once I’ve been able to make the repo public and you can take a look.
HI @elizabethlorenc,
We are very interested in trying this!
We will watch this thread for updates.
Thanks!
Hi everyone,
The voxalign repo is now public at GitHub - brown-bnc/voxalign: automated prescription of MRS voxels. Please let me know if you have questions, feedback, etc.! Happy to help any interested groups give it a try.
Elizabeth