CORRELATION OF LIVER DENSITY BY MAGNETIC-RESONANCE-IMAGING AND HEPATIC IRON LEVELS - A NONINVASIVE MEANS TO EXCLUDE HOMOZYGOUS HEMOCHROMATOSIS

The diagnosis of hemochromatosis requires liver biopsy and the quantif ication of hepatic iron. Magnetic resonance imaging (MRI) of the liver shows a characteristic decrease in tissue signal intensity in iron ov erload states, but its role in the diagnosis of hemochromatosis has no t been fully delineated. Forty-three patients (31 men and 12 women) we re referred for the evaluation of hemochromatosis based upon a fasting transferrin saturation >55% and/or a serum ferritin >400 ng/ml in mal es or >300 ng/ml in females. Each patient prospectively underwent MRI of the liver prior to percutaneous liver biopsy and quantitative hepat ic iron determination. Homozygous hemochromatosis was diagnosed in 10 patients based upon an hepatic iron/age index greater than or equal to 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 m s, TE = 30 ms). Signal intensity values were blindly determined for re gions of interest in liver and skeletal muscle at T1 and proton densit y. Ratios of liver to muscle (LM) for T1 and proton density (PD) calcu lated from these values showed a significant correlation with quantita tive iron by multiple regression analysis. The LMPD ratio provided the best correlation with hepatic iron (r = -0.6946; p < 0.001). Linear r egression analysis also provides an equation that can be used to predi ct hepatic iron based upon the LMPD ratio; mu g/g of hepatic iron = (- 5,174 x LMPD) + 9,932. All patients with LMPD ratios of >0.5 had hepat ic iron/age indices of <2.0, thereby excluding homozygous hemochromato sis. These results suggest that LMPD ratios derived from MRI of the li ver can accurately predict hepatic iron content. These ratios can be c linically useful in the evaluation of hemochromatosis among patients w ho either refuse or have contraindications to liver biopsy.