Return To Abstract Listing

Gastrointestinal Imaging

E2840. Differentiating Steatosis from Non-Alcoholic Steatohepatitis (NASH) with Ultrasound

Nelson S,  Schwantes J,  Chaudhuri J. San Antonio Uniformed Services Health Education Consortium , San Antonio, Texas

Address correspondence to J. Schwantes (

Objective: Non-alcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in the Western world. NAFLD describes a continuum of hepatic pathologies ranging from steatosis, non-alcoholic steatohepatitis (NASH) to cirrhosis. Differentiating Steatosis from NASH is important as patients with NASH may progress to cirrhosis, which in turn can also increase risk of hepatocellular carcinoma (HCC). Steatosis and possibly NASH may be reversible. Liver biopsy is the current gold standard for diagnosis. Recently, the Ultrasonographic Fatty Liver Indicator (US-FLI) was developed to identify patients with NASH. It is comprised of sonographic features including liver kidney contrast, posterior attenuation, vessel blurring, difficulty visualizing the gallbladder wall, difficulty visualizing the diaphragm, and areas of focal fatty sparing. It is scored on a scale of 2-8 with a minimum score of =2 equaling NAFLD. Using this method, we proposed to validate this tool in the obese population (average BMI 31) and hypothesized that one or more sonographic features would be independent predictors of NASH.

Materials and Methods: A retrospective analysis of 208 patients was performed. All patients underwent biopsy with diagnosis and staging (Brunt score). There were 14 normal livers, 89 with steatosis and 105 with NASH. Exclusion criteria included: alcohol use greater than 21 units/week for men, 14 units/week for women, diagnosis of viral hepatitis, hemochromatosis, Wilson’s disease or cirrhosis. Statistical analysis using R-statistical software package (R core team, 2016) was performed. For non-normally distributed variables, the Wilcoxon-Rank sum test was performed. Sensitivities, Specifities, Positive Predictive Values (PPV) and Negative Predictive Values (NPV) were calculated with contingency tables. Logistic regression was performed to identify independent predictors of NASH.

Results: Only two sonographic indicators were statistically different in the Steatosis vs. NASH group, Gallbladder wall visualization and Vessel blurring (p = 0.01). In an analysis of the pathologic stage (I) of NASH versus Steatosis both Gallbladder wall visualization and Vessel blurring showed a statistical difference between Steatosis and stage I NASH (p = 0.01). Logistic regression revealed that only Vessel blurring was an independent predictor of NASH (p = 0.01). In contrast, only Gallbladder wall visualization was specific for NASH (89%). Several indicators were sensitive for NASH including Vessel blurring (93%) and Posterior attenuation (86%). A US-FLI score < 4 was highly suggestive of lack of NASH (NPV of 88%).

Conclusion: Our results overall correlate well with previously published data suggesting this tool is also valid in the obese population. Vessel blurring was the best individual predictor of NASH in this retrospective analysis while gallbladder wall visualization was specific for NASH. Correlation with laboratory testing may further strengthen this tool and enhance the ability to distinguish Steatosis from NASH and is an ongoing component of this research.