Correlation of liver density by magnetic resonance imaging and hepatic iron levels. A noninvasive means to exclude homozygous hemochromatosis

Abstract

The diagnosis of hemochromatosis requires liver biopsy and the quantification of hepatic iron. Magnetic resonance imaging (MRI) of the liver shows a characteristic decrease in tissue signal intensity in iron overload states, but its role in the diagnosis of hemochromatosis has not been fully delineated. Forty-three patients (31 men and 12 women) were referred for the evaluation of hemochromatosis based upon a fasting transferrin saturation > 55% and/or a serum ferritin > 400 ng/ml in males or > 300 ng/ml in females. Each patient prospectively underwent MRI of the liver prior to percutaneous liver biopsy and quantitative hepatic iron determination. Homozygous hemochromatosis was diagnosed in 10 patients based upon an hepatic iron/age index > or = 2. MRI was performed with a 1.5 Tesla system using standard spin-echo sequences (T1; TR = 300-500 ms, TE = 13-17 ms, PD; TR = 2,000-2,600 ms, TE = 30 ms). Signal intensity values were blindly determined for regions of interest in liver and skeletal muscle at T1 and proton density. Ratios of liver to muscle (LM) for T1 and proton density (PD) calculated from these values showed a significant correlation with quantitative iron by multiple regression analysis. The LMPD ratio provided the best correlation with hepatic iron (r = -0.6946; p < 0.001). Linear regression analysis also provides an equation that can be used to predict hepatic iron based upon the LMPD ratio; micrograms/g of hepatic iron = (-5,174 x LMPD) + 9,932. All patients with LMPD ratios of > 0.5 had hepatic iron/age indices of < 2.0, thereby excluding homozygous hemochromatosis. These results suggest that LMPD ratios derived from MRI of the liver can accurately predict hepatic iron content. These ratios can be clinically useful in the evaluation of hemochromatosis among patients who either refuse or have contraindications to liver biopsy.