Heterogeneity of intrahepatic fat distribution determined by (18)F-FDG PET and CT

Hepatic steatosis is common but less is known of the heterogeneity of hepatic fat distribution and its clinical significance. Our objective was to measure the regional variabilities within the liver of standardised uptake values (SUV) as potential markers of hepatic fat distribution heterogeneity.

Twenty-four patients having routine, clinically indicated PET/CT with (18)F-FDG and a wide range of fatty liver severity were selected. Maximum SUV (SUVmax), average SUV (SUVave), both calculated using lean body mass, and CT density were measured in 12 × 2-cm diameter ROIs in each patient. SUVave was also measured over the left ventricular cavity (SUVLV). Mean values of SUV indices, their ratios with SUVLV, and CT density in the 12 ROIs were calculated. Regional variabilities of SUV indices were expressed as coefficients of variation (CV; standard deviation/mean). Body mass index (BMI) was estimated from height and body weight, and %body fat and lean body mass from height, weight and gender.

Mean SUVmax/SUVave correlated significantly with mean CT density (r = -0.51; p < 0.02). In contrast, mean SUVmax, mean SUVave and their ratios with SUVLV showed no correlation with CT density. Mean CT density correlated with weight (r = -0.59; p < 0.005), BMI (r = -0.57; p < 0.01) and %body fat (r = -0.49; p < 0.02). Corresponding correlation coefficients for mean SUVmax/SUVave were 0.74 (p < 0.001), 0.65 (p < 0.001) and 0.46 (p < 0.03). In contrast, mean SUVmax, mean SUVave and their ratios with SUVLV showed no correlation with BMI, weight and %body fat. The CV of SUVmax/SUVave (r = -0.67; p < 0.001), but not the CVs of SUVmax or SUVave, correlated with mean CT density.

SUVmax/SUVave and CT density are markers of hepatic steatosis. The regional variability of SUVmax/SUVave may be a marker of hepatic fat distribution heterogeneity. The novel concept is introduced that hepatic fat distribution heterogeneity may be a marker of hepatic pathology and of clinical value, and deserves further exploration with texture analysis.