Introduction: We investigated the clinical usefulness of liver MRI in a large multicenter study because the control of iron concentration (LIC) is crucial in thalassemia major (TM) prognosis and conflicting results have been reported in small and heterogeneous groups of patients using Magnetic Resonance Imaging (MRI), the only very common non-invasive procedure.
Material and methods: One hundred and eight consecutive TM patients, selected according to a specific protocol, were included in the study. In 29 of these LIC was measured by atomic absorption spectrophotometry on liver biopsy and expressed as microgram/gr dry weight. MRI was performed with a 0.5 T superconducting unit. The ratio of the average signal intensity (SIR) of the liver to that of paraspinal muscle was calculated and expressed as a percentage.
Results: Mean +/- (SD) of SIR was 100 +/- 26% (test for normality: p = 0.02) and of LIC was 3677 +/- 4662 micrograms/g/dry liver (test for normality: p < 0.00001). The logit transformation of LIC (lLIC) achieved both normality of distribution of lLIC and linear regression of lLIC on SIR (t = 7.36; p < 0.00001) according to the equation: Y = -0.0136 + SIR* (0.157) = lLIC log10 (K-LIC/LIC) where K = 79.433. Thus the values of LIC are expressed by the equation LIC K/(1 + 10 y) micrograms/g/dry liver. Moreover, a mild correlation was found between SIR and AST (p = 0.01; r = -0.30), ALT (p = 0.02; r = -0.21), gamma GT (p = < 0.01; r = -0.37; r = -0.25).
Conclusions: This study shows that LIC can be calculated as a function of SIR. A validation study is necessary before introducing this prediction rule in clinical practice.