Fatty infiltration of the liver: quantification with phase-contrast MR imaging at 1.5 T vs biopsy.

Quantification of hepatic fat content by application of MR phase-contrast imaging (Dixon method) at 1.5 T was compared with results of biopsy in 16 patients with a variety of liver abnormalities. Motion artifact was suppressed by employing six or eight averages of short TR in-phase (echo offset, 0 msec), out-of-phase (echo offset, 1.1 msec), and in-phase (echo offset, 2.2 msec) spin-echo pulse sequences. The 360 degree out-of-phase sequence was used to assess the impact of T2* decay on this method of estimating fat fraction. A standard two-echo long TR sequence also was obtained in all patients. Histologic preparations from the biopsy specimens were examined by a pathologist who had no knowledge of the MR results and were graded according to overall visual assessment as belonging to one of four categories of fat fraction. Results of the MR-calculated apparent fat fraction were compared directly with biopsy category and were also placed in MR fat fraction categories, allowing estimation of the statistical correlation between the biopsy and MR grading systems. Eight of eight patients with biopsy categories indicating a fat fraction of less than 0.25 were computed by MR to have a fat fraction of less than 0.1. Seven of eight patients with biopsy categories indicating a fat fraction of greater than 0.25 were computed by MR to have a fat fraction of at least 0.24. The MR-calculated apparent fat fraction category correlated significantly with the histologic biopsy category (r = .86, p less than .01). When compared with the in-phase image, decreased signal from liver was visually apparent on the 180 degree out-of-phase images in all cases in which the fat fraction was at least 0.24, but there was no indication of fatty liver on the standard T1- or T2-weighted images. Calculated T2 also showed no correlation with degree of fatty deposition. Correction for T2* decay by using the 360 degree out-of-phase acquisition in addition to the standard 0 degree and 180 degree out-of-phase images had little effect on fat fraction computation. Phase-contrast MR is a promising noninvasive method for quantitative assessment of fatty deposition in the liver.