Discordance diagnosis between B-mode ultrasonography and MRI proton density fat fraction for fatty liver

Abstract

We aimed to evaluate the frequency and causes of discordant results in fatty liver (FL) diagnosis between B-mode ultrasonography (B-USG) and magnetic resonance imaging proton density fat fraction (MRI-PDFF). We analyzed patients who underwent both B-USG and MRI-PDFF within a 6-month interval. We made a confusion matrix for FL diagnosis between B-USG and MRI-PDFF and identified four discordant groups as follows: (1) the "UFL-MnFL-wo" group [B-USG FL-MRI-PDFF no FL without chronic liver disease (CLD) or liver cirrhosis (LC)]; (2) the "UFL-MnFL-w" group (B-USG FL-MRI-PDFF no FL with CLD or LC); (3) the "UnFL-MFL-wo" group (B-USG no FL-MRI-PDFF FL without CLD or LC); and (4) the "UnFL-MFL-w" group (B-USG no FL-MRI-PDFF FL with CLD or LC). We compared the "UFL-MnFL-wo" group with the control group in terms of various parameters. We found 201 patients (201/1514, 13.3%) with discordant results for FL diagnosis between B-USG and MRI-PDFF. The "UFL-MnFL-wo" group accounted for the largest portion at 6.8% (103/1514), followed by the "UFL-MnFL-w" group (79/1514, 5.2%) and the "UnFL-MFL-w" group (16/1514, 1.1%). The mean and right PDFF values, body mass index, and abdominal wall thickness were significantly higher in the "UFL-MnFL-wo" group than in the control group (p <= 0.001). The frequency of discordant results in the diagnosis of FL between B-USG and MRI-PDFF could be identified. The causes of discordant results were that B-USG was fairly accurate in diagnosing FL disease and that accompanying CLD or LC hindered the evaluation of FL.