Aim: The increasing prevalence of fatty liver disease requires routine assessment methods. Proton magnetic resonance spectroscopy ((1)H MRS) is increasingly used for steatosis measurement, but due to cost, is unlikely to become a widely-used screening tool. Ultrasound is cheaper and more widely available, although subject to observer variability. Our aim was to determine the sensitivity and specificity of ultrasound against (1)H MRS, using MRS as a gold standard, for the detection and quantification of hepatic fat content.
Methods: Fifty adults participated (43 men, seven women) in this study. Hepatic steatosis was assessed by ultrasound and (1)H MRS. Images were graded by two independent radiologists to classify severity and distribution of liver fat.
Results: Ultrasound detected liver fat infiltration in 82% of cases measurable by (1)H MRS, while liver fat was detectable in 44% of cases graded absent by ultrasound. Ultrasound grading was subjective, with the radiologists in agreement in 53% of cases (kappa = 0.39, P = 0.002). Considerable overlap in intrahepatocellular lipid content was observed between different grades: absent (0.0-1.58%), mild (2.2-16.2%), moderate (4.9-26.7%) and severe (8.1-76.8%) steatosis. Ultrasound could not detect liver fat levels below 2% as measured by (1)H MRS Conclusion: Ultrasound is less sensitive than (1)H MRS in detecting very low levels of liver fat content, but is sensitive to fatty infiltration greater than 2%. There is a tendency of higher ultrasound grades to correlate with higher degrees of fatty infiltration, although some overlap exists. Our findings are still consistent with ultrasound being useful as a low cost screening tool.