PURPOSE: To assess the role of magnetic resonance (MR) imaging in detection and quantification of liver iron overload. MATERIALS AND METHODS: MR imaging at 0.5 T was prospectively performed on 77 patients (67 with liver iron overload and 10 without) who underwent a liver biopsy with biochemical determination of the liver iron concentration (LIC) (normal, < 36 mumol per gram of liver tissue dry weight). Ratios of signal intensities and liver T2 relaxation time were calculated from images obtained with spin-echo and breath-hold gradient-echo (GRE) sequences. RESULTS: Liver-to-tissue signal intensity ratios were better correlated with LIC than T2 relaxation time. Long-echo-time GRE sequences were the most sensitive for detection of slight overload. Thus, high sensitivity (94%) and specificity (90%) were obtained with a liver-to-fat ratio threshold of 1. The quantification of iron with MR imaging was accurate when the LIC was 80-300 mumol/g. For heavy overload, above 300 mumol/g, quantification was impossible owing to complete signal loss. Pancreatic and splenic signal intensity were unchanged in most cases. CONCLUSION: This method, which can be improved by using more sensitive sequences with a high-field-strength system, should be competitive with biopsy for the diagnosis of substantial liver iron overload.
Gandon et al. (Tue,) studied this question.