Comparing Visual, Morphometric and Biochemical Methods for Estimating Liver Fat Content
M Li, J Song, S Mirkov, MJ Ratain, W Liu, J Hart. Univ of Chicago, Chicago, IL
Background: The degree of macrovesicular steatosis of liver is usually assessed by visual estimation of fat content on histology. It is known that visual estimation is subject to intra- and inter-observer variations. More objective methods, such as morphometric and biochemical methods, have also been described. We conducted the first study to evaluate the relationship among these three methods.
Design: We obtained 26 fresh frozen liver specimens from the Pharmacogenetics of Anti-cancer Agent Research Group at University of Chicago. Each specimen was divided into four 50 mg aliquots. Three were homogenized and treated by organic solvents to extract fat, and fat content was measured as the average weight percentage. The other was formalin fixed for H&E histology. Two pathologists independently reviewed the slides and estimated fat content. The slides were then scanned at 100x, and the percentage of area occupied by fat was measured by morphometry using ImagePro Plus 6.0. For morphometry, distinct fat droplets were used as references for image segmentation, and artifacts were minimized by appropriate thresholds for object size, shape and roundness.
Results: The fat content measured by the 3 methods showed different ranges: visual, 0-80%; morphometry, 0-19.6%; biochemical, 2.2-15%. Very strong correlation was seen between any two methods (correlation coefficients 0.83-0.95, all p values < 0.01). Linear regression equations and correlation coefficients are shown in Figure 1.
Conclusions: Visual, morphometric and biochemical methods all showed very strong correlation, suggesting that all three are accurate and reliable for estimating fat content. The morphometric and biochemical methods provide results as a continuous variable and may be more useful in research settings. In the clinical setting, visual estimation has been proven to be reliable and correlate well with the severity of steatosis; therefore, morphometric and biochemical method may not add more value. The consistently higher percentage given by the visual method reflects the fact that in this method the percentage of cells with fat is counted, rather than the area occupied by fat droplets (morphometric method) or the total fat content (biochemical method).
Category: Liver & Pancreas
Monday, March 22, 2010 9:30 AM
Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 212, Monday Morning