Multiparametric magnetic resonance for the non-invasive diagnosis of liver disease.

The study by Banerjee et al. (2014) addresses the pressing need for reliable, non-invasive methods to diagnose and stage liver diseases, given the limitations of liver biopsy, which include invasiveness, sampling errors, and observer variability. 

In this prospective study, 79 adult patients scheduled for liver biopsy underwent a novel multiparametric magnetic resonance (MR) imaging protocol. This protocol combined T1 mapping, proton magnetic resonance spectroscopy, and T2* mapping to quantify liver fibrosis, steatosis, and haemosiderosis, respectively. 

The MR imaging results showed strong correlations with histological findings:

• Fibrosis: Spearman’s correlation coefficient (rₛ) = 0.68, p < 0.0001

• Steatosis: rₛ = 0.89, p < 0.001

• Haemosiderosis: rₛ = -0.69, p < 0.0001

The diagnostic accuracy, assessed by the area under the receiver operating characteristic curve (AUROC), was high for all three conditions:

• Fibrosis: AUROC = 0.94

• Steatosis: AUROC = 0.93

• Haemosiderosis: AUROC = 0.94

These findings suggest that the multiparametric MR imaging technique provides a comprehensive, accurate, and non-invasive assessment of liver pathology. It holds promise as a safer alternative to liver biopsy, particularly for detecting early-stage liver diseases.