Non‐invasive evaluation of hepatic fibrosis using magnetic resonance and ultrasound techniques

In his recent commentary, Pinzani discussed approaches to non‐invasive evaluation of liver fibrosis and addressed some of the methods currently under investigation (Gut 2006;55:310–12). We agree that liver histology is a surrogate end point and Pinzani emphasised the need for longitudinal studies based on hard clinical endpoints. However, he states “At present, CT and MR can indicate the presence of cirrhosis with high specificity but with very low sensitivity”. In reply, we wish to counter this statement and would like to present data from emerging technologies, including magnetic resonance spectroscopy (MRS), ultrashort echo time (UTE) MR imaging (MRI), and microbubble ultrasound.

We used 31P MRS in vivo to characterise hepatic fibrosis in chronic hepatitis C (CHC) infection.¹ The phosphomonoester to phosphodiester (PME/PDE) ratio provided an index of cell membrane turnover and was found to correlate closely with disease severity, assessed by liver histology (Ishak system). A PME/PDE ratio ⩾0.3 provided a sensitivity and specificity of 82% and 81%, respectively, for the diagnosis of cirrhosis, comparable with many indirect serological markers. There was a monotonic increase in PME/PDE ratio with increasing disease activity and statistically significant differences between mild hepatitis, moderate/severe hepatitis, and cirrhosis.¹ We also demonstrated the potential utility of UTE MRI² and showed the relaxation time, T2*, was significantly different between controls and patients with cirrhosis. Functionally decompensated liver disease (Child's grade C) differed significantly from functionally compensated liver disease (Child's A/B). Significant differences in diffusion weighting MRI indices between patients with cirrhosis and normal volunteers have also been reported.³

Pinzani highlighted the controversy surrounding usage of Doppler ultrasonography (US) parameters. The study he cited of Doppler‐US indexes, in conjunction with clinical signs and biochemical measures, should be interpreted with caution as the diagnostic accuracy of Doppler‐US variables on an intention to diagnose basis was 87%, based on a protocol in which three variables were assessed in a stepwise manner.⁴ However, in contrast, our analysis of Doppler‐US in the assessment of CHC fibrosis demonstrated no significant difference between Doppler indexes with increasing severity of liver disease.⁵

It should be noted that US microbubble contrast agents have also been employed to evaluate liver fibrosis, through hepatic vein transit time (HVTT) measurements.⁶,⁷ HVTT decreases with increasing severity of liver disease, due to associated circulatory changes, which include arterialisation of the hepatic sinusoidal bed, the presence of intrahepatic and intrapulmonary shunting, and the hyperdynamic circulation present in patients with cirrhosis.⁶ An HVTT of less than 24 seconds was 100% sensitive and 96% specific for diagnosis of cirrhosis.⁶,⁸ In a cohort of 85 CHC patients, HVTT demonstrated 100% sensitivity and 80% specificity for cirrhosis, and 95% sensitivity and 86% specificity for differentiation of mild hepatitis from more severe liver disease.⁷ There was also a significant difference between moderate/severe hepatitis and these groups. Such a clear difference between Ishak grades of disease severity may, perhaps, inform treatment decisions.

For a non‐invasive biomarker of liver disease to be employed in a clinical setting, its place in clinical practice must be appraised. In the context of CHC, a role in the decision to start antiviral therapy and in monitoring of treatment response would be important. Biomarkers may also have a role in the stratification of patients with cirrhosis according to the risk of clinical outcomes, such as variceal bleeding and the development of hepatocellular carcinoma, as has been suggested for transient elastography by Foucher and colleagues.⁹ While it is ideal to find a single biomarker, a profile of tests embracing different modalities, including imaging and serological markers, is more realistic and will probably add value to clinical management algorithms.