Curious: MR spectroscopy question

[-Goose-]

---

Members don't see this ad.

I usually check out the case of the day on auntminnie.com when I am bored at work (not that I know anything about radiology at this point-- its just interesting). When I was reading today's (5/31) case, I came across something that was interesting:

http://education.auntminnie.com/QMachine.ASP?UID=1KJ0SO66&PageId=3&Sess=0061457


Is MR spectro. within the usual scope of a radiologist or of a pathologist (or both)? It was pretty interesting to see, regardless.

 

[ariwax]

MR spec is under the aegis of radiology, and is a somewhat rarified corner of MRI that has been around for a while. I only ever saw it in neuro, and it's basically a method for looking at a single voxel (= volume pixel) at figuring out the molecular composition based on magnetic resonance. This can clue you in on whether a lesion is tumor, necrosis, normal brain tissue, etc.

Standard mass spectrometry, on the other hand, is a method used by pathologists and other lab-types to determine molecular composition of a sample. It's done not with magnetic resonance, but by breaking down a sample into component ions and analyzing them by mass-to-charge (m/z) ratio, then usually cross-referencing that fingerprint electronically to a database of proteomics. It's a totally different process, involving in some cases (i.e., matrix-assisted laser desorption-ionization, or MALDI) the use of a chemical printer synced up with a laser, the use of a vacuum chamber, and so on. It's a powerful modality that is actually really hot right now.

 

[Koil Gugliemi]

Research places that are equiped with newer magnets can actually do some imaging based on some of these metabolites. But that's not ready for primetime. Most academic neurorad places will do a fair bit of single-voxel MR spects as in the AM COTD.

 

[ariwax]

MR imaging based on choline et al? How does that work?

 

[Koil Gugliemi]

It's complicated. See Magnetic Resonance in Medicine 46:1072–1078 (2001).

http://www3.interscience.wiley.com/cgi-bin/fulltext/88011770/HTMLSTART

It's also being used quite a bit in prostate CA and breast CA research.

 

[Neuronix]

One can use a chemical shift imaging (CSI) sequence to obtain spectra from a number of voxels in an area, either in 2D or 3D. There are a lot of different ways to do this, but most are used in research only. The Siemens clinical scanners I work with do include CSI sequences, which are phase encoded to obtain each spectrum from each voxel. Regardless of the sequence details (I can get more for you if you wish), the output is a map of spectra overlayed on an anatomical image. The neat thing is that you can now integrate the area under the peaks corresponding to the chemical shifts you're interested in to create a metabolite "image". By showing the integrated areas of the peaks overlaid on the anatomical image, usually in color form to make it look neato, distributions of metabolites are seen in the brain.

The problem is that most in vivo NMR spectra are averaged hundreds of times with a fairly high TR (repetition time) on the order of at least a second. This means to get a "high quality" spectrum for single voxel spectroscopy you're looking at somewhere on the order of 5-10 minutes (at least). In CSI, you are often only getting one voxel average per excitation. This means you either have to run the thing a really long time to get many averages, or just accept relatively low quality spectra.

 

[Uncle_Tbag]

Much like you, I don't know much about MR Spectroscopy, but all of the MRS scans done at our clinic (and it isn't many) are read by one of our neurologists.