QSM: what acceleration type and normalization should you use for GE scanners?

Hi.
I have some questions regarding setup of QSM protocol for GE scanners.

First, in the recent consensus it is mentioned that you should use ASSET (similar to SENSE).

  • however this may be due to the fact that only ASSET was supported for SWAN, since software version MR30 and now ARC (similar to GRAPPA ) and hypersense (PI-CS for GE) is supported. I wonder which acceleration scheme works best, theortetically or empirically.

Second, it is recommende to use normalization methods such as use prescan-normalize (Siemens) or CLEAR/SYNERGY (Philips). However PURE (GE) is not mentioned. Is there any drawbacks on using this method? I read in a paper that N4 correction should be avoided for phase images (SCENIC for GE). In the manual of the scanner, there is a sentence that PURE should be used with care in dynamic scans such as DCE. I checked on the scanner that the real and imaginary values as well as mangitude values are affected. Should PURE be used? If not used the magnitude images are so biased that they end up resulting in poor brain masks.

Thanks in advance for your reply!