A prospective comparison of three ultrasound-based techniques in quantitative diagnosis of hepatic steatosis in NAFLD.

Purpose: To investigate the correlation between different ultrasound attenuation-based techniques and to compare their diagnostic performances using proton magnetic resonance spectroscopy (H-MRS) as a reference standard.

Methods: Participants who had clinical suspicion of nonalcoholic fatty liver disease (NAFLD) were prospectively recruited. Each subject had ultrasound with attenuation imaging (ATI) or quantitative ultrasound including tissue attenuation imaging (TAI) and tissue scatter-distribution imaging (TSI), and controlled-attenuation parameter (CAP) and H-MRS if available. The technical success rates, intra-observer repeatabilities of attenuation and backscattering coefficient were evaluated. ATI, TAI and CAP were three attenuation-based techniques. Spearman coefficient was used to test correlations among them and H-MRS. In addition, the diagnostic performances of these parameters for detecting ≥ 5% or 10% hepatic steatosis were evaluated.

Results: 130 participants had ultrasound scanning. Among them, 67 had CAP and 48 had H-MRS. The technical success rates were all 100%. The intra-observer repeatabilities of them were also excellent (ICCs > 0.90) and AC-ATI correlated well with AC-TAI (r = 0.752). AC-ATI, AC-TAI showed moderate correlation with CAP, (r = 0.623, 95% CI 0.446-0.752, P < 0.001; r = 0.573, 95% CI 0.377-0.720, P < 0.001). For correlation with H-MRS, ATI and TAI performed better than CAP(r = 0.587; r = 0.712; r  = 0.485). The AUCs of ATI, TAI, TSI and CAP for detecting ≥ 5% hepatic steatosis were 0.883, 0.862, 0.870 and 0.868, respectively. The AUC improved to 0.907 when TAI and TSI were combined (P < 0.05). When detecting ≥ 10% hepatic steatosis, the AUCs were 0.855, 0.702, 0.822 and 0.838, respectively.

Conclusion: Different ultrasound attenuation-based techniques were well correlated and exhibited good diagnostic performances in quantitative diagnosis of hepatic steatosis, however, the threshold values were different. Combinations of multiple parameters may improve the diagnostic performance in detecting hepatic steatosis.

Trial Registration: The study has been registered online ( https://www.chictr.org.cn ; unique identifier: ChiCTR2300069459).