Diagnostic accuracy of ultrasound-derived fat fraction for the detection and quantification of hepatic steatosis in patients with liver biopsy.

Abstract

Purpose: This retrospective study was conducted to investigate the diagnostic accuracy of ultrasound-derived fat fraction (UDFF) for grading hepatic steatosis using liver histology as the reference standard.

Methods: Seventy-three patients with liver disease were assessed using UDFF and liver biopsy. Pearson's test and the Bland-Altman plot were used to assess the correlation between UDFF and histological fat content in liver sections. The UDFF cutoff values for histologically proven steatosis grades were determined using the area under the receiver operating characteristic curve (AUROC).

Results: The median age of the patients was 66 (interquartile range 54-74) years, and 33 (45%) were females. The UDFF values showed a stepwise increase with increasing steatosis grade (p < .001) and were strongly correlated with the histological fat content (r = .7736, p < .001). The Bland-Altman plot revealed a mean bias of 2.384% (95% limit of agreement, - 6.582 to 11.351%) between them. Univariate regression analysis revealed no significant predictors of divergence. The AUROCs for distinguishing steatosis grades of ≥ 1, ≥2, and 3 were 0.956 (95% confidence interval [CI], 0.910-1.00), 0.926 (95% CI, 0.860-0.993), and 0.971 (95% CI, 0.929-1.000), respectively. The UDFF cutoff value of > 6% had a sensitivity and specificity of 94.8% and 82.3%, respectively, for diagnosing steatosis grade ≥ 1. There was no association between UDFF and the fibrosis stage.

Conclusion: UDFF shows strong agreement with the histological fat content and excellent diagnostic accuracy for grading steatosis. UDFF is a promising tool for detecting and quantifying hepatic steatosis in clinical practice.