Liver iron concentration quantification by MRI: are recommended protocols accurate enough for clinical practice?

Area of Science:

• Medical Imaging
• Hepatology
• Biomarkers

Background:

• Accurate quantification of liver iron concentration (LIC) is crucial for managing iron overload disorders.
• Magnetic Resonance Imaging (MRI) offers a non-invasive method for assessing LIC.
• The Rennes University (URennes) algorithm is one method used for MRI-based LIC quantification.

Purpose of the Study:

• To evaluate the diagnostic accuracy of the URennes algorithm for quantifying LIC using MRI.
• To compare MRI-derived LIC measurements with liver biopsy results.

Main Methods:

• A cohort of 171 patients underwent MRI-based LIC quantification using the URennes model between 1999 and 2006.
• LIC measurements were compared against results from liver biopsy, considered the gold standard.
• Statistical analysis included correlation coefficients, diagnostic accuracy, sensitivity, specificity, and predictive values.

Main Results:

• The URennes method demonstrated a strong correlation with biopsy (r=0.86) but showed a tendency to overestimate iron overload.
• Diagnostic accuracy was 61.4%, with significant misclassification rates for no overload and moderate overload.
• High sensitivity (92.3%) for high overload and high negative predictive value (100%) for absence of haemochromatosis were observed.

Conclusions:

• The URennes MRI method is useful for ruling out or diagnosing high iron overload in 74.3% of patients.
• Overestimation of iron overload by the URennes algorithm limits its overall diagnostic performance.
• Further refinement may be needed to improve the accuracy of LIC quantification in intermediate cases.