Masoud Shiehmorteza MD, Abstract Co-Author: Nothing to Disclose

Irene Mwangi MD, Abstract Co-Author: Nothing to Disclose

Mark Bydder PhD, Abstract Co-Author: Research grant, General Electric Company

Takeshi Yokoo MD, Abstract Co-Author: Nothing to Disclose

Tanya Wolfson MS, Abstract Co-Author: Nothing to Disclose

Claude B. Sirlin MD, Presenter: Research grant, Bayer AG, West Haven, CT Research grant, General Electric Company, Fairfield, CT Research grant, Bracco Group Contract, Isis Pharmaceuticals, Inc Contract, Wyeth Speaker Bureau, Bayer AG Consultant, Bayer AG

Anthony Gamst PhD, Abstract Co-Author: Nothing to Disclose

Daniel Fox, Abstract Co-Author: Nothing to Disclose

Julie Collins, Abstract Co-Author: Nothing to Disclose

00030490-DMT et al, Abstract Co-Author: Nothing to Disclose

To assess accuracy, repeatability, and robustness of a multi-echo gradient-recalled-echo (GRE) MR imaging for in vivo quantification of hepatic fat in human subjects

In this HIPAA-compliant, IRB-approved study on subjects with known or suspected fatty liver disease, hepatic fat fraction (FF) was calculated using T1 independent, T2* corrected 6-echo GRE sequence, LIPO-Quant (Liver Imaging of Phase related signal Oscillation and Quantification), modeling fat as 5-peak spectrum. To assess LIPO-Quant accuracy, we reviewed 314 MRIs ( 154 at 1.5T, 160 at 3T) that included LIPO-Quant and, as the reference standard, single-voxel T2-corrected MR STEAM spectroscopy. LIPO-Quant and co-localized MRS FFs were correlated. To assess robustness, LIPO-Quant sequences were performed at 1.5 and 3T in 55 subjects while modifying several imaging parameters (slice thickness, slice gap, matrix, BW, TR). Standard deviation of FF in each subject across parameters was calculated. To assess measurement-remeasurement variability, LIPO-Quant sequences were repeated three times using identical parameters on 100 MRIs. To assess exam-reexam variability, 28 subjects underwent two separate LIPO-Quant sequences 60-90 minutes apart. Bland-Altman repeatability of hepatic FF measurements, obtained from 3 ROIs, was calculated.  

 LIPO-Quant FF correlated with MRS FF (regression slope, 0.98; intercept, 0.62 percentage points). Differences in FF measurements between MRS and LIPO-Quant on Bland-Altman plot were small (mean, -0.38; ±2 SD of the difference, -3.3 to 2.6 percentage points). For robustness assessment, SD of FF across all LIPO-Quant sequences per subject ranged from 0.1 to 2.5 percentage points (median 0.7 percentage points). The SD of the FF measurement for assessing repeatability was 0.1 to 1.8 percentage points (median of 0.5 percentage points). Differences in FF measurements between first and second examinations in  exam-reexam repeatability were small (mean difference, -0.07 percentage points; ±2 SD of the difference, -1.6 to 1.4 percentage points).

LIPO-Quant, a multi-echo gradient-recalled-echo (GRE) MR imaging sequence, is an accurate, repeatable, and robust technique for in vivo quantification of hepatic fat fraction in human subjects.

An accurate, repeatable, and robust non-invasive MR imaging method for in vivo quantification of hepatic fat content

Shiehmorteza, M, Mwangi, I, Bydder, M, Yokoo, T, Wolfson, T, Sirlin, C, Gamst, A, Fox, D, Collins, J, et al, 0, Accuracy, Repeatability, and Robustness of MR Hepatic Fat Quantification by Using a Multi-Echo Gradient-Recalled-Echo (GRE) MR Imaging in Human Subjects.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8016502.html