[Standardization of the quantification of iron concentration in the liver by magnetic resonance imaging].

OBJECTIVE To calibrate 1.5 tesla magnetic resonance scanners for the quantification of the concentration of iron in the liver. MATERIAL AND METHODS We analyzed twenty-eight 1.5 tesla magnetic resonance scanners using a phantom with four tubes containing different concentrations of iron (III) chloride and one tube without iron. The phantom represented two typical patients: one with moderate iron overload and one with high iron overload. We measured the signal intensity ratio between each iron-containing tube and the tube without iron; then we calculated the theoretical levels of iron concentration in each scanner according to the model for the two levels of overload. We compared the results of each scanner with those of the reference scanner in which the model and the phantom had been designed, and we calculated the percentage of difference between the two scanners. RESULTS The mean difference in the ratios compared to the reference center was 11% (0.3-39). The mean concentration of iron was 71 μmol Fe/g for moderate overload and 193 μmol Fe/g for high overload. The mean difference was 6% (1.2- 7%) and 3.4% (0-16%). respectively. In two scanners, we applied a correction factor so that the difference was below 25% in all cases. CONCLUSION We calibrated twenty-eight 1.5 tesla scanners for the concentration of iron in the liver and achieved variability less than 25%.