MRSI water acquisition

Any suggestions how to properly and efficiently acquire water reference data for MRSI? For particulars - using standard stock Siemens PRESS MRSI. Typical acquisition is 2 averages and takes about 15min. When I simply reduce to 1 avg and turn water suppression off, it’s a 7min acquisition… I’m hoping this can be done more efficiently?


You could drop the TR (and with that the excitation flip angle). This will then make the signal T1 dependent and would need correcting (which cannot be done perfectly), but water has the signal to support it.

Are the options for different k-space sampling available, usually something like “full”, “elliptical”, “circular”. The latter ones miss off the corners of k-space and will reduce the duration, without much detrimental effect on the point spread function.

Thanks, great suggestion about the k-space sampling (or “Phase encoding” as Siemens calls it), that saves a nice 1/4 scan time!

I’m debating whether interpolation for water signal would be a good idea or not. Effectively water signal would be acquired for every 4 “voxels” and interpolated in between.

By reducing the resolution by four times? I’d be wary of that because the point spread function will end up very broad. You are really open to some nasty effects arising from the ripple of the PSF when the resolution gets too low. I would definitely look at TR first.

Alright, got it, bad idea, thank you!

TR for acquisition is 1700ms, minimum achievable is 1510ms. But water acquisition with elliptical k-space sampling is under 5 min, so I think that’s tolerable.