MR elastography of healthy liver parenchyma: Normal value and reliability of the liver stiffness value measurement

Abstract

To investigate the normal liver stiffness value using MR elastography (MRE) and to compare the repeatability and reproducibility of three measurement methods. Forty-nine subjects who underwent liver MR imaging including elastography using a 1.5 Tesla scanner, who had normal laboratory test results, and who underwent liver donation, were included in this study. Two, independent readers measured the liver stiffness value of the hepatic parenchyma using three methods, including evaluating various liver parenchyma portions using different-sized, regions-of-interest (ROIs): 2cm-ROI-per-slice in the right lobe (2cm-per-slice); 1cm-ROI-per-segment (1cm-S); and 70%-of-the-surface area (70% S). The mean liver stiffness values were compared between gender using the t-test and among the age groups using one-way analysis of variance. The reproducibility and repeatability of the three liver stiffness value measurement methods were determined using intraclass correlation coefficients (ICCs) and the 95% Bland-Altman limits of agreement. The liver stiffness values in living liver donors ranged from 1.54 to 2.87 kPa. The mean stiffness value was 2.05 kilopascal using the 2cm-per-slice method, 2.01 kilopascal with the 1cm-S method and 2.12 kilopascal using the 70% S method. There was no significant difference in the liver stiffness value according to the gender or age factors. For reproducibility, the ICCs were 0.416 with the 2cm-per-slice method, 0.800 using the 1cm-S method, and 0.845 with the 70% S method. In terms of repeatability, the ICCs were 0.238 using the 2cm-per-slice method, 0.914 with the 1cm-S method, and 0.852 using the 70% S method. The ICCs determined using the 2cm-per-slice method were significantly lower than those of the 1cm-S or 70% S method for assessing both reproducibility and repeatability. The 95% limit of agreement was 54.0% with the 2cm-per-slice method, 24.0% using the 1cm-S method, and 18.8% with the 70% S method, in terms of reproducibility. To assess the repeatability, the 95% limit of agreement was 63.3% using the 2cm-per-slice method, 20.6% with the 1cm-S method, and 17.4% using the 70% S method. The mean liver stiffness values in living donors ranged from 2.05 to 2.12 kilopascals and did not differ significantly for either gender or age. The 1cm-S and 70% S methods were significantly more reliable compared with the 2cm-per-slice method.