Validation of a B-mode ultrasonography scoring system for assessing liver steatosis: A comparison with MRI-Derived proton density fat fraction

Aim: Noninvasive detection of liver steatosis and monitoring its progression are essential for effective therapeutic management. We validated the diagnostic performance of an ultrasonography (US)-based steatotic liver scoring system, derived from the B-mode method and the hepatic steatosis index (HSI) for the detection of liver steatosis, as identified by proton density fat fraction (PDFF) measurements on magnetic resonance imaging (MRI). Methods: A total of 916 patients with chronic liver disease were included in the analysis. Results: The median MRI-PDFF value was 4.0% (interquartile range: 2.0-9.7). The distribution of scores (0/1/2/3/4/5/6) according to the US-based steatotic liver scoring system was as follows: 475, 36, 99, 78, 141, 66, and 21, respectively. The median HSI was 32.8 (28.9-37.4). A significant positive association between advancing scores of the US-based steatotic liver scoring system and increasing MRI-PDFF values was observed (p < 0.001). The diagnostic performance of the US-based steatotic liver scoring system and HSI for detecting steatosis grades >= 1, >= 2, and 3, as determined by MRI-PDFF, showed areas under the receiver operating characteristic curve of 0.940 and 0.842 (p < 0.001), 0.949 and 0.847 (p < 0.001), and 0.945 and 0.864 (p < 0.001), respectively. When the cutoff values of the scoring system were set at 2, 3, and 4 for steatosis grades >= 1, >= 2, and 3, the sensitivity and specificity were 90.6% and 90.5%, 92.9% and 83.0%, and 93.3% and 84.0%, respectively. Conclusion: The B-mode US-based steatotic liver scoring system demonstrated robust diagnostic capability in detecting liver steatosis.