MR Spectroscopy-derived Proton Density Fat Fraction Is Superior to Controlled Attenuation Parameter for Detecting and Grading Hepatic Steatosis

Radiology. 2018 Feb;286(2):547-556. doi: 10.1148/radiol.2017162931. Epub 2017 Sep 15.

Abstract

Purpose To prospectively compare the diagnostic accuracy of controlled attenuation parameter (CAP) obtained with transient elastography and proton density fat fraction (PDFF) obtained with proton magnetic resonance (MR) spectroscopy with results of liver biopsy in a cohort of adult patients suspected of having nonalcoholic fatty liver disease (NAFLD). Materials and Methods The institutional review board approved this study. Informed consent was obtained from all patients. The authors evaluated 55 patients suspected of having NAFLD (40 men, 15 women). Patients had a median age of 52.3 years (interquartile range [IQR], 43.7-57.6 years) and a median body mass index of 27.8 kg/m2 (IQR, 26.0-33.1 kg/m2). CAP and PDFF measurements were obtained on the same day, within 27 days of biopsy (IQR, 7-44 days). CAP and PDFF were compared between steatosis grades by using the Jonckheere-Terpstra test. Diagnostic accuracies of CAP and PDFF for grading steatosis were assessed with receiver operating characteristic (ROC) analysis. Within-weeks reproducibility (CAP and PDFF) and within-session repeatability were assessed with linear regression analyses, intraclass correlation coefficients, and coefficients of variation. Results Steatosis grades at liver biopsy were distributed as follows: S0, five patients; S1, 24 patients; S2, 17 patients; and S3, nine patients. Both PDFF and CAP helped detect histologically proven steatosis (≥S1), but PDFF showed better diagnostic accuracy than CAP in terms of the area under the ROC curve (0.99 vs 0.77, respectively; P = .0334). PDFF, but not CAP, enabled the grading of steatosis (P < .0001). For within-weeks reproducibility, the intraclass correlation coefficient with PDFF was higher than that with CAP (0.95 vs 0.65, respectively; P = .0015); coefficients of variation were similar (19% vs 11%, P = .55). Within-session repeatability of CAP was good, with a coefficient of variation of 4.5%. Conclusion MR spectroscopy-derived PDFF is superior to CAP in detecting and grading liver steatosis in human NAFLD. © RSNA, 2017 Online supplemental material is available for this article.

Publication types

  • Comparative Study
  • Evaluation Study
  • Multicenter Study

MeSH terms

  • Adipose Tissue / pathology*
  • Adult
  • Biopsy
  • Elasticity Imaging Techniques / standards
  • Female
  • Humans
  • Liver / pathology*
  • Magnetic Resonance Imaging / standards
  • Male
  • Middle Aged
  • Non-alcoholic Fatty Liver Disease / pathology*
  • Proton Magnetic Resonance Spectroscopy / standards
  • ROC Curve