Liver fat quantitation at 3T using a single breath hold multi-echo sequence

Purpose: Magnetic resonance imaging and spectroscopy provide a tool for fat quantitation for a wide range of liver conditions such as steatosis, non alcoholic fatty liver disease (NAFL), or non alcoholic steatohepatitis (NASH). Various methods are being used for fat quantitation, such as MR spectroscopy or 3-point Dixon. A practical method based on dual flip angle and in-phase and out-of-phase (IP-OP) echoes has been developed on 1.5T [1] and validated at 3T [2]. The method consisted of 3 distinct scans: a high flip angle (70°) and low flip angle (20°) in- and out-of-phase gradient echo sequences, followed by another T2 weighted dual-echo gradient echo scan for T2* measurement. Here we present an improvement of the method, where we acquired 6 echoes (3 in-phase/out-of-phase pairs) with dynamically varying flip angle, providing fat quantitation and T2* correction in a single breath hold scan.

Materials and Methods: The pre-clinical evaluation of the technique was performed in 6 volunteers. Images were acquired on a 3T clinical MR system (Philips Achieva, R2.5) with a 6-channel phased-array surface coil. The exam consisted of a localizer, a multi-echo dual flip angle gradient-echo breath hold scan, and a single voxel spectroscopy scan. The sequence parameters for the multi-gradient-echo sequence were following: multi-slice 2D fast field multi-echo (mFFE), 6 echoes, TR/TE 184/1.15ms, ΔTE 1.2ms, dynamic flip angle: 70°/20°, voxel size 2.2×2.47mm, 17 slices of 7mm, 1mm gap, FOV 360×280mm (106×164 matrix), SENSEx1.8, 1135Hz/pix bandwidth, 24s breath hold duration. Proton MR spectroscopy was acquired with a single voxel PRESS sequence, TR 4000ms, TE 45, 65ms, 2 NSA, 2 phase cycles, 1.95Hz spectral resolution, 30×30×30mm³ voxel size placed in a homogenous region of the liver, 16s scan duration. Fat content was calculated from the m-FFE images using the dual flip angle technique. Integration of water and fat spectral peaks at TE=45ms and 65ms allowed for removal of T2 bias in MRS fat estimation.

Results: By implementing the dynamic flip angle into the multi-echo gradient echo sequence, we were able to obtain high and low T1 contrast, in and out of phase echoes in a singe breath hold scan, allowing for fat content measurement with T2* correction. Fat content values were calculated by the dual flip angle multi-echo method and MR spectroscopy. Color -coded fat content maps and T2* maps were created from the dual flip angle multi-echo gradient-echo images. Good correlation was found between the dual flip-angle MR imaging and MR spectroscopy (R=0.95).

Conclusion: The dual flip angle method has been developed and validated as a reliable and accurate fat content measurement at 1.5T and 3T. While demonstrating good correlation with the standard MRS technique, this method provides better spatial coverage within comparable scan time. While MRS is considered a gold standard, like histopathology it measures only a small volume in the liver which does not yield a measurement of a heterogeneous disease. In this work we further improved its implementation, by providing fat content and T2* measurement of the whole liver in a single breath hold scan.