Chulalongkorn Medical Journal


Background: The diagnosis of hepatic steatosis has been well-described on unenhanced computed tomography (UECT) study. However, current UECT often discarded from the abdominal CT protocol due to radiation dose reduction. Objective: To determine accuracy of contrast-enhanced CT (CECT) for diagnosis of hepatic steatosis. Methods: A total of 1,001 patients who underwent unenhanced and portal venous phase abdominal CT studies were assessed by using region-of-interests of liver and spleen, and visual detection of focal fat sparing. UECT diagnostic criteria were used as the standard reference. Results: The optimal cut-offs on CECT images were 110 Hounsfield unit (HU) of liver attenuation and -20 HU of liver-splenic differential (L-S) attenuation. Sensitivity, specificity, accuracy and receiver operating characteristic curve areas for quantitative liver attenuation values were 90.4%, 73.6%, 75.9% and 0.926, respectively; and for L-S attenuation values were 83.1%, 70.8%, 72.5% and 0.856, respectively. Qualitatively, geographic fat sparing was 100.0% specificity; however, its sensitivity (50.7%) was rather low. Conclusion: Portal venous phase CECT can be used for detection of hepatic steatosis in abdominal CECT study without the preceding unenhanced phase.


Faculty of Medicine, Chulalongkorn University



First Page


Last Page