CT Attenuation Thresholds for Grading Hepatic Steatosis in Adults with Overweight or Obesity and Metabolic Dysfunction-Associated Steatotic Liver Disease.

Background/

Objectives

To derive computed tomography (CT) attenuation thresholds for grading hepatic steatosis and evaluate their agreement with ultrasonography (US) in adults with overweight or obesity and metabolic dysfunction-associated steatotic liver disease (MASLD) from southeastern Mexico.

Methods

Adults (≥18 years) with MASLD who underwent both CT and US were included. Hepatic steatosis on CT was defined by liver attenuation < 40 Hounsfield units (HU) or a liver-to-spleen attenuation difference > 10 HU. US steatosis was graded using a three-point echogenicity scale. CT attenuation thresholds were derived in a training dataset (70%) using sequential receiver operating characteristic (ROC) analyses and the Youden index and were subsequently evaluated in an internal hold-out dataset (30%). Histopathological findings were available in a subgroup for additional evaluation. Agreement between CT-predicted and US-based steatosis grades was assessed using Spearman's rank correlation coefficient (ρ) and quadratic-weighted Cohen's kappa (κ).

Results

In total, 105 participants were included. The training-set ROC analyses identified CT attenuation thresholds of ≤28.15 HU for Grade 3 steatosis (AUC = 0.995), 28.15-35.50 HU for Grade 2 (AUC = 0.983), and >35.50 HU for Grade 1. In the internal hold-out dataset, CT-predicted and US-based steatosis grades showed a strong positive correlation (ρ = 0.946, p < 0.001) and almost perfect agreement (quadratic-weighted κ = 0.940, p < 0.001). Liver biopsy data were available for 16 participants, and CT-predicted steatosis grade was positively associated with histological fat percentage (ρ = 0.595, p = 0.015).

Conclusions

The proposed CT attenuation thresholds showed high agreement with US-based steatosis grading, and their association with histopathological fat content provided additional supportive evidence in a subgroup of participants. These findings suggest that CT attenuation may provide a useful quantitative approach to grading hepatic steatosis and could contribute to more consistent steatosis assessment across different healthcare settings.