Quantification of the fat fraction in the liver using dual-energy computed tomography and multimaterial decomposition.

Abstract

The objective of this study was to evaluate the feasibility and accuracy of postcontrast, dual-energy computed tomography (DECT) using multimaterial decomposition (MMD) for quantification of the hepatic fat fraction (HFF). Various degrees of fatty liver were produced in 16 rabbits divided into 4 groups, by feeding them a high-fat, high-cholesterol diet. After 6 weeks, precontrast, single-energy computed tomography (SECT) and postcontrast DECT were performed, and HFF maps were obtained from the DECT using MMD. T2*-corrected, multiecho, chemical-shift imaging for obtaining the proton density fat fraction (PDFF) map and histologic assessment of the hepatic steatosis were also performed. The Pearson correlations coefficients were calculated, and Bland-Altman analysis was performed of the fat fractions determined by DECT, magnetic resonance imaging (MRI), and pathology. Receiver operating characteristic curve analysis was performed to detect 5% hepatic steatosis or greater. The histologic HFF was used as the reference standard, and PDFFs determined by chemical-shift imaging as the technical standard. The HFFs calculated by MMD of postcontrast DECT scans were strongly correlated with those of pathology as well as the PDFFs of MRI (P < 0.05), and the mean differences of the Bland-Altman plots between the HFFs of DECT with MMD and the PDFFs of MRI were close to zero. The diagnostic performance of postcontrast DECT with MMD for detecting 5% hepatic steatosis or greater was comparable to that of the attenuation value of precontrast SECT and the PDFFs of MRI (P = 0.17-0.82). Quantification of the HFF using postcontrast DECT with MMD was feasible and showed comparable sensitivity and specificity to the precontrast SECT for the detection of hepatic steatosis.