Original Article
The influence of liver fat deposition on the quantification of the liver-iron fraction using fast-kilovolt-peak switching dual-energy CT imaging and material decomposition technique: an in vitro experimental study
Abstract
Background: To assess the feasibility of dual-energy spectral computed tomography (DECT) for quantifying the liver iron content (LIC) with material decomposition (MD) technique in vitro.
Methods: Liver-iron mixture samples (model A) and liver-iron-fat mixture samples (model B) were prepare and scanned by a single source DECT using GSI mode with successive tube currents of 200, 320, and 485 mA. A standard algorithm of 1.25 mm was used to reconstruct iron (fat) MD images and iron (water) MD images. The iron concentrations of all samples were measured and analyzed by Spearman’s rank correlation and linear regression analysis.
Results: Significant positive linear correlations were found between virtual iron content (VIC) and LIC in the absence of fat (model A) and in the presence of fat (model B) in the range of LIC 0 to 25 mg/mL. The lines of best fit to model A had slopes around 1.1 and an intercept around (−1.5) mg/mL for iron (water) MD images, and had slopes around 1.1 and an intercept around (−10) mg/mL for iron (fat) MD images. The lines of best fit to the model B had slopes around 1.5 and an intercept around (−15) mg/mL. At the same value of LIC (LIC >0), the VIC values of model A were always higher than those of model B. At the high value of LIC (12.5 mg/mL), the VIC values of model B were similar, but they differed greatly from those of model A.
Conclusions: The fast-kilovolt-peak switching dual-energy CT imaging and MD techniques allow for quantification of iron content. Fat and the post-reconstruction algorithm of iron (fat) MD images, were confounding factors, and led to the underestimation and overestimation of LIC, respectively.
Methods: Liver-iron mixture samples (model A) and liver-iron-fat mixture samples (model B) were prepare and scanned by a single source DECT using GSI mode with successive tube currents of 200, 320, and 485 mA. A standard algorithm of 1.25 mm was used to reconstruct iron (fat) MD images and iron (water) MD images. The iron concentrations of all samples were measured and analyzed by Spearman’s rank correlation and linear regression analysis.
Results: Significant positive linear correlations were found between virtual iron content (VIC) and LIC in the absence of fat (model A) and in the presence of fat (model B) in the range of LIC 0 to 25 mg/mL. The lines of best fit to model A had slopes around 1.1 and an intercept around (−1.5) mg/mL for iron (water) MD images, and had slopes around 1.1 and an intercept around (−10) mg/mL for iron (fat) MD images. The lines of best fit to the model B had slopes around 1.5 and an intercept around (−15) mg/mL. At the same value of LIC (LIC >0), the VIC values of model A were always higher than those of model B. At the high value of LIC (12.5 mg/mL), the VIC values of model B were similar, but they differed greatly from those of model A.
Conclusions: The fast-kilovolt-peak switching dual-energy CT imaging and MD techniques allow for quantification of iron content. Fat and the post-reconstruction algorithm of iron (fat) MD images, were confounding factors, and led to the underestimation and overestimation of LIC, respectively.
