Quantification of liver fat content: comparison of triple-echo chemical shift gradient-echo imaging and in vivo proton MR spectroscopy

Abstract

Purpose: To validate a triple-echo gradient-echo sequence for measuring the fat content of the liver, by using hydrogen 1((1)H) magnetic resonance (MR) spectroscopy as the reference standard.

Materials and methods: This prospective study was approved by the appropriate ethics committee, and written informed consent was obtained from all patients. In 37 patients with type 2 diabetes (31 men, six women; mean age, 56 years), 3.0-T single-voxel point-resolved (1)H MR spectroscopy of the liver (Couinaud segment VII) was performed to calculate the liver fat fraction from the water (4.7 ppm) and methylene (1.3 ppm) peaks, corrected for T1 and T2 decay. Liver fat fraction was also computed from triple-echo (consecutive in-phase, opposed-phase, and in-phase echo times) breath-hold spoiled gradient-echo sequence (flip angle, 20 degrees), by estimating T2* and relative signal intensity loss between in- and opposed-phase values, corrected for T2* decay. Pearson correlation coefficient, Bland-Altman 95% limit of agreement, and Lin concordance coefficient were calculated.

Results: Mean fat fractions calculated from the triple-echo sequence and (1)H MR spectroscopy were 10% (range, 0.7%-35.6%) and 9.7% (range, 0.2%-34.1%), respectively. Mean T2* time was 14.7 msec (range, 5.4-25.4 msec). Pearson correlation coefficient was 0.989 (P < .0001) and Lin concordance coefficient was 0.988 (P < .0001). With the Bland-Altman method, all data points were within the limits of agreement.

Conclusion: A breath-hold triple-echo gradient-echo sequence with a low flip angle and correction for T2* decay is accurate for quantifying fat in segment VII of the liver. Given its excellent correlation and concordance with (1)H MR spectroscopy, this triple-echo sequence could replace (1)H MR spectroscopy in longitudinal studies.