The accuracy of sonography in predicting steatosis and fibrosis in chronic hepatitis C.

The accuracy and clinical significance of sonography (US) in demonstrating fatty liver and hepatic fibrosis in chronic hepatitis C (CHC) are rarely reported. US had sensitivity 71.1%, specificity 72.9%, 58.7% positive predictive value (PPV), and 82.3% negative predictive value (NPV) in demonstrating histological steatosis > or =5%. US had sensitivity 85.7%, specificity 60.4%, 13% PPV, and 98.4% NPV in demonstrating histological steatosis > or =30% with clinical significance in predicting prognosis and therapeutic response in CHC. Subjects with fatty liver on US had a greater prevalence of body mass index (BMI) > or =25 kg/m2, inflammation-necrosis grade >2, and total bilirubin <1.2 mg/dl in multivariate analyses. US had sensitivity 27.4%, specificity 62.5%, 71.9% PPV, and 19.7% NPV in demonstrating histological fibrosis of stage II or above, and sensitivity 13.6%, specificity 66.3%, 9.4% PPV, and 75.0% NPV in demonstrating fibrosis of stage III or above. There was no correlation between fibrotic sonographic patterns and histological stage of fibrosis (r = -0.167, P = 0.083). Besides hepatic steatosis, clinicians should be alert to the possibility of advanced necrosis-inflammation grade in interpreting a report of bright liver on gray-scale US. Gray-scale US cannot replace liver biopsy as the optimal diagnostic procedure for the prediction of hepatic steatosis and fibrosis prior to initiating therapy for CHC.