• liver;
  • steatosis;
  • living donor;
  • plain CT;
  • chemical-shift MR



To compare plain computed tomography (CT) and chemical-shift MR imaging (CSI) for establishing a noninvasive method to estimate the degree of steatosis.

Materials and Methods

A total of 58 patients who had histological proof of liver tissue and underwent CT and/or CSI within two weeks, were studied. They consisted of 38 living donor candidates and 20 liver surgery patients without hepatitis or cirrhosis. The CT index (CTI) and MR index (MRI) were defined and correlated to histological grades of steatosis. Correlation was also performed exclusively for mild steatosis (grades 0–2).


Both plain CT (ρ = 0.74, P < 0.0001) and CSI (ρ = 0.83, P < 0.0001) had good correlation with histological grades of steatosis, showing no significant difference between them. For mild steatosis, CSI (ρ = 0.77, P < 0.0001) had better correlation than CT (ρ = 0.49, P = 0.001) (P = 0.049). Accuracy in differentiation between grade 0–1 vs. grade 2–4 steatosis was 85% using a CTI >6, and 91% using an MRI <0.03 as criteria. That between grade 0–2 and grade 3–4 was 91% using a CTI >–8 and 88% using an MRI <0.25 as criteria.


Both plain CT and CSI were useful in estimating the degree of steatosis but CSI was superior to CT in differentiating mild steatosis. These indices may at least partially replace percutaneous biopsy in evaluating grades of steatosis for living donors. J. Magn. Reson. Imaging 2008;28:678–684. © 2008 Wiley-Liss, Inc.