Arthur Peter Wunderlich PhD, Presenter: Consultant, Siemens AG

Stefan Andreas Schmidt, Abstract Co-Author: Nothing to Disclose

Meinrad Johannes Beer MD, Abstract Co-Author: Research Consultant, Shire plc

Holger Cario, Abstract Co-Author: Nothing to Disclose

Volker Rasche MD, PhD, Abstract Co-Author: Nothing to Disclose

Annihilation of remaining magnetization by radio frequency (RF) spoiling alters image contrast compared to solely gradient spoiling. We studied influence of RF-spoiling (RFS) on liver iron content (LIC) results of gradient echo (GRE) MRI utilizing signal intensity ratio (SIR), i.e. ratio of liver to muscle signal.

In liver iron overloaded patients, a spin-echo based MRI method yielded LIC reference values. Under approval of our local ethics committee, GRE data was also acquired with TE 4.76 and 9.53 ms, TR 120 ms, and FA 20° and 90°. RF spoiling was randomly switched on (50:50). Axial slices were obtained with 5 mm thickness, FoV of 380 mm and 2x2 mm resolution. 12 patients suitable for SIR analysis, i.e. with reference LIC below 350 mmol/kg liver dry weight, were randomized for each group. Additionally, 3 patients with LIC < 350 mmol/kg scanned with both +/- RFS were selected, summing up to a total of 27 patients and 30 investigations. ROIs were drawn in liver and paraspinal muscles, and analyzed using a method proposed by Gandon. Results were correlated to LIC reference by linear regression analysis seperately for +/- RFS.  

LIC determined by SIR correlated well with reference LIC for both + and - RFS. R2 was larger without RFS (0.85 vs. 0.71 +RFS). Differences depend on LIC: for LIC below 150 mmol/kg, data +RFS show higher values compared to data -RFS, and vice versa in patients with LIC above 150 mmol/kg.

The popular GRE based method for LIC quantification proposed by Gandon et al. relies on liver to muscle SIR. Contrast changes depending on the spoiling scheme have be considered, since differences in correlation to reference LIC value were observed when comparing RF-spoiled vs. non-RF-spoiled GRE. This is important in case MRI for LIC determination is performed on systems from different vendors since similar acronyms are often used for different GE techniques. We observed a difference in linearity between data acquired with and without RF spoiling. Further studies are needed since differences are most prominent at LIC values occurring only rarely in our patient cohort. The superior linearity of reference LIC and results of non-RF-spoiled GRE suggests avoiding RF-spoiling for public available SIR based LIC determination.  

Choice of different spoiling schemes in SIR/GRE-based LIC determination may impact results of LIC quantification.

Wunderlich, A, Schmidt, S, Beer, M, Cario, H, Rasche, V, MRI Based Liver Iron Content Determination Using Signal Intensity Ratio Analysis: RF Spoiled vs. Not RF Spoiled Gradient Echo.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14045629.html