Utility and cost evaluation of multiparametric magnetic resonance imaging for the assessment of non-alcoholic fatty liver disease

Background

Validated diagnostic tools that are accurate, cost effective and acceptable to patients are required for disease stratification and monitoring in NAFLD.

Aims

To investigate the performance and cost of multiparametric MRI alongside existing biomarkers in the assessment of NAFLD.

Methods

Adult patients undergoing standard of care liver biopsy for NAFLD were prospectively recruited at two UK liver centres and underwent multiparametric MRI, blood sampling and transient elastography withing 2 weeks of liver biopsy. Non-invasive markers were compared to histology as the gold standard.

Results

Data were obtained in 50 patients and 6 healthy volunteers. Corrected T1 (cT1) correlated with NAFLD activity score (ρ = 0.514, P < .001). cT1, enhanced liver fibrosis (ELF) test and liver stiffness differentiated patients with simple steatosis and NASH with AUROC (95% CI) of 0.69 (0.50-0.88), 0.87 (0.77-0.79) and 0.82 (0.70-0.94) respectively and healthy volunteers from patients with AUROC (95% CI) of 0.93 (0.86-1.00), 0.81 (0.69-0.92) and 0.89 (0.77-1.00) respectively. For the risk stratification of NAFLD, multiparametric MRI could save £150,218 per 1000 patients compared to biopsy. Multiparametric MRI did not discriminate between individual histological fibrosis stages in this population (P = .068).

Conclusions

Multiparametric MRI accurately identified patients with steatosis, stratifies those with NASH or simple steatosis and reliably excludes clinically significant liver disease with superior negative predictive value (83.3%) to liver stiffness (42.9%) and ELF (57.1%). For the risk stratification of NAFLD, multiparametric MRI was cost effective and, combined with transient elastography, had the lowest cost per correct diagnosis.