Liver-spleen contrast standardization of gadolinium-ethoxybenzyl-diethylenetriamine penta-acetic acid-enhanced magnetic resonance imaging based on cross-calibration

BACKGROUND: Liver-spleen contrast in the hepatobiliary phase highly depends on the devices used for liver function tests. This study aimed to develop and validate a method to convert liver-spleen contrast data acquired with another device to reference liver-spleen contrast data, using the regression line of phantom contrasts (i.e., cross-calibration). METHODS: As cohort studies, two-dimensional gradient echo images of T1-weighted fat-suppression in the hepatobiliary phase were retrospectively obtained and analyzed for a total of 126 patients who underwent gadolinium-ethoxybenzyl-diethylenetriamine penta-acetic acid-enhanced magnetic resonance imaging using four different magnetic resonance imaging scanner-coil combinations. The liver-spleen contrast measured from these images was converted into reference liver-spleen contrast using cross-calibration with purified water and gadolinium-ethoxybenzyl-diethylenetriamine penta-acetic acid phantoms of 0.06, 0.14, 0.27, 0.63, 1.37, and 2.82 mM/L. At this point, the error of the regression lines with phantom contrasts was assessed and corrected. Lastly, the liver-spleen and converted liver-spleen contrasts, which are the values before and after cross-calibration respectively, were compared with reference liver-spleen contrast in three cases using different coils and magnetic resonance imaging scanners from a reference device. RESULTS: Regarding the regression lines with phantom contrasts, the coefficient of determination was 0.99. Although regression lines with phantom contrasts tended to be 0.0105 lower in logarithmic contrasts than those with liver-spleen contrast, no significant difference was observed between the two lines (P=0.0612) by analysis of covariance. In the case of different coils, there was a significant difference between liver-spleen and reference liver-spleen contrasts (P<0.00001), but there was no significant difference between converted liver-spleen and reference liver-spleen contrasts (P=0.492). Moreover, the regression equation between converted liver-spleen and reference liver-spleen contrasts corresponded with an identity line. Likewise, in the two cases of different magnetic resonance imaging scanners, there was a significant difference between liver-spleen and reference liver-spleen contrast (both P<0.00001), but there was no significant difference between converted liver-spleen and reference liver-spleen contrast (P=0.923 and P=0.541). CONCLUSIONS: Cross-calibration using the precision and valid regression lines with phantom contrasts had high accuracy and utility.