Aims: The aims of the study were to compare the efficacy of magnetic resonance imaging (MRI), biochemical assay and histological grading in estimating liver iron content, and to evaluate the value of liver to muscle signal intensity ratio (L/M ratio) on spin-echo T1-weighted images in this role.
Materials and methods: Thirty-nine homozygous beta-thalassaemics had their L/M ratio measured on MRI, followed by ultrasound-guided liver biopsies with histological grading of iron storage and biochemical quantification of liver iron concentration (LIC-b) using atomic absorption spectrophotometry.
Results: A significant difference in L/M ratios between the four grades of iron storage on histology was observed (P < 0.001). The coefficient of correlation was -0.67 between L/M ratio and LIC-b ranging from 2 to 21.6 mg/g dry weight. Specific values of L/M ratio reliably reflected liver iron content at clinically important levels. A L/M ratio of < 0.6 has 86% sensitivity and 100% specificity in the prediction of grade 3 or 4 iron storage histologically and 81% sensitivity and 81% specificity in predicting LIC-b > 15 mg/g. A L/M ratio of > 0.8 predicts a histological iron storage grading of 0 or 1 with a 100% sensitivity and 74% specificity.
Conclusion: L/M ratio on MRI is of value as a non-invasive alternative to repeated liver biopsies for estimating liver iron content at clinically important thresholds.
Copyright 2001 The Royal College of Radiologists.