Clinical value of liver ultrasound for the diagnosis of nonalcoholic fatty liver disease in overweight and obese patients

Background & Aims Liver ultrasound (US) is usually used in the clinical setting for the diagnosis and follow-up of patients with nonalcoholic fatty liver disease (NAFLD). However, no large study has carefully assessed its performance using a semiquantitative ultrasonographic scoring system in overweight/obese patients, in comparison to magnetic resonance spectroscopy (1H-MRS) and histology. Methods We recruited 146 patients and performed: a liver US using a 5-parameter scoring system, a liver 1H-MRS to quantify liver fat content, and a liver biopsy to assess histology. All measurements were repeated in a subgroup of patients (n = 62) after 18 months of follow-up. Results The performance of liver US (parenchymal echo alone) was rather modest, and significantly worse than 1H-MRS (AUROC: 0.82 [0.69–0.94] vs. 0.96 [0.90–1.00]; P = 0.04). However, the AUROC improved when different echographic parameters were taken into account (AUROC: 0.89 [0.83–0.96], P = 0.15 against 1H-MRS). Optimum sensitivity for liver US was achieved at a liver fat content ≥12.5%, suggesting that below this threshold, liver US is less sensitive. Liver 1H-MRS showed a high accuracy for the diagnosis of NAFLD, and correlated strongly with histological steatosis (r = 0.73, P