Accuracy of a commercial high-speed T2-corrected multi-echo single voxel sequence at 3T as a function of R2* for measurement of liver fat content

Abstract

Commercial methods for single breath-hold proton density fat fraction (PDFF) quantification in liver can suffer from bias due to the presence of iron. In this work, the PDFF accuracy of a high-speed T2-corrected multi-echo (HISTO) sequence was evaluated at 3T in phantoms at variable R2*, mimicking different iron levels. PDFF errors up to 70% for R2* larger than 150 s-1 were obtained, suggesting that HISTO is unreliable at large liver R2*, as seen with moderate to high iron overload. A Dixon sequence was more accurate at large R2*. Both techniques agreed in patients and phantoms at low PDFF and R2*.