Reproducibility of Intra- and Interscanner Measurements of Liver Fat Using Complex Confounder-corrected Chemical Shift Encoded MRI at 3.0 Tesla

The purpose of this study was to prospectively evaluate the reproducibility of the proton density fatfraction (PDFF) of the liver using the IDEAL algorithm, a quantitative confounder-corrected chemicalshift-encoded MRI method. Data were obtained from 15 volunteers on four different days. The first and the third examinations were conducted on scanner one with one-week intervals, while the second and the fourth tests were performed on scanner two with same time interval. For each test, two MR scans were performed, one before and one after a meal. Regions-of-interest measurements were manually calculated to estimate the PDFF in the right and left lobes on the PDFF images. Reproducibility was measured using the intra-class correlation coefficient (ICC). The ICCs of the PDFF in the right and left lobes were 0.935 and 0.878, respectively. The intra-scanner ICCs of the right lobe before and after a meal or at a one-week interval were 0.924 and 0.953, respectively. The inter-scanner ICCs of PDFF the next day and at a one-week interval were 0.920 and 0.864, respectively. The PDFF of liver derived from IDEAL demonstrated high intra-and inter-scanner measurement reproducibility. The PDFF of the right lobe before a meal was more reproducible than after-meal measurements.