Correlation between ultrasonography and MR proton density fat fraction techniques in evaluating the severity of liver steatosis

Abstract

Background and Aim: To investigate the relationship between ultrasonography (US) and magnetic resonance (MR) proton density fat fraction (PDFF) techniques, using the modified DIXON method, in determining the severity of liver steatosis. Materials and Methods: This study included seventy consecutive patients who underwent upper abdominal MRI for various reasons between June 2016 and January 2017. Fatty liver staging was performed using US as indicated.The liver fat percentage was measured and staged according to PDFF values. Results: In the study, of the 70 cases, 36 were male and 34 were female. On US, 18.5% of the cases had stage 0, 32.8% had stage 1, 42.8% had stage 2, and 5.7% had stage 3 liver steatosis. A significant correlation was found between ultrasonographic evaluation and PDFF in determining the percentage of liver fat (r=0.775, p<0.001). When comparing the percentages, MR-evaluated PDFF and ultrasonographic staging were most compatible at grade 3 and least compatible at grade 2. When the PDFF threshold value was set at 8.1%, the sensitivity of US in distinguishing between obvious and indistinct steatosis was 97.1%, and the specificity was 88.9%. Conclusion: Ultrasound continues to be a useful tool for detecting fatty liver disease. However, magnetic resonance (MR) proton density fat fraction (PDFF) imaging is essential for accurately determining the severity and prevalence of steatosis. Our study revealed inconsistencies between US and MR PDFF in grading liver steatosis, showing higher agreement in severe cases and lower agreement in moderate cases. Therefore, we recommend classifying steatosis as either uncertain or apparent rather than using a grading system in US.