Diagnostic performance of controlled attenuation parameter for grading hepatic steatosis in MASLD: an...

This study aimed to evaluate the diagnostic accuracy of the controlled attenuation parameter (CAP) for grading hepatic steatosis in metabolic dysfunction-associated steatotic liver disease (MASLD), using magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) as the non-invasive reference standard.

This single-center retrospective study included 120 participants (17 healthy controls and 103 patients with MASLD) who underwent MRI-PDFF, transient elastography (TE), and laboratory testing within predefined time intervals. Participants were stratified by MRI-PDFF values into no (S0, n = 17), mild (S1, n = 20), moderate (S2, n = 59), or severe (S3, n = 24) steatosis. Group comparisons were performed using ANOVA or Kruskal-Wallis tests, correlations were assessed with Pearson or Spearman coefficients, and independent factors associated with CAP were identified by multivariate linear regression. The diagnostic performance of CAP was evaluated using receiver operating characteristic (ROC) curve analysis. Bootstrap resampling (2000 iterations) was used to estimate the 95% confidence intervals (CIs) for the optimal CAP cut-off values for each steatosis grade.

CAP values increased progressively with steatosis severity (S3 > S2 > S1 > S0; all p < 0.001). Hepatic steatosis grade was independently associated with CAP values after adjusting for confounders (p < 0.001). CAP showed good diagnostic performance for identifying ≥S1 (area under the curve [AUC] = 0.924) and ≥S2 steatosis (AUC = 0.947), and acceptable performance for identifying S3 steatosis (AUC = 0.837). The optimal CAP cut-offs were 239 dB/m (95% CI: 235.5-240.0) for ≥S1, 278 dB/m (95% CI: 273.0-307.5) for ≥S2, and 314 dB/m (95% CI: 303.5-357.5) for S3, with corresponding sensitivities of 95.0, 94.9, and 91.7%, and specificities of 88.2, 83.8, and 62.5%.

CAP showed good diagnostic performance for grading hepatic steatosis in MASLD and may serve as a practical non-invasive tool for steatosis assessment, particularly in resource-limited settings where MRI-PDFF is not readily available.