Normal Liver Stiffness Measured with MR Elastography in Children

Abstract

Background Stiffness thresholds for liver MR elastography in children vary between studies and may differ from thresholds in adults. Normative liver stiffness data are needed to optimize diagnostic thresholds for children. Purpose To determine normal liver stiffness, and associated normal ranges for children, as measured with MR elastography across vendors and field strengths. Materials and Methods This was a prospective multicenter cohort study (ClinicalTrials.gov identifier: NCT03235414). Volunteers aged 7-17.9 years without a known history of liver disease were recruited at four sites for a research MRI and blood draw between February 2018 and October 2019. MRI was performed on three vendor platforms and at two field strengths (1.5 T and 3.0 T). All MRI scans were centrally analyzed; stiffness, proton density fat fraction (PDFF), and R2* values were expressed as means of means. Mean and 95% confidence intervals (CIs) for liver stiffness were calculated. Pearson correlation coefficient (r), two-sample t test, or analysis of variance was used to assess univariable associations. Results Seventy-one volunteers had complete data and no documented exclusion criterion (median age, 12 years; interquartile range [IQR], 10-15 years; 39 female participants). Median body mass index percentile was 54% (IQR, 32.5%-69.5%). Mean liver stiffness was 2.1 kPa (95% CI: 2.0, 2.2 kPa) with mean ± 1.96 kPa standard deviation of 1.5-2.8 kPa. Median liver PDFF was 2.0% (IQR, 1.7%-2.6%). There was no association between liver stiffness and any patient variable or MRI scanner factor. Conclusion Mean liver stiffness measured with MR elastography in children without liver disease was 2.1 kPa (similar to that in adults). The 95th percentile of normal liver stiffness was 2.8 kPa. Liver stiffness was independent of sex, age, or body mass index and did not vary with MRI scanner vendor or field strength. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Yin in this issue.

Figures

Figure 1:

Study sample flow diagram. Boxes on right denote excluded participants based on established exclusion criteria or missing data. BMI = body mass index, LFTs = liver function tests, PDFF = liver proton density fat fraction, %ile = percentile.

Figure 2a:

(a) Histogram and (b) Tukey box/dot plot show distribution of liver stiffness values measured with MR elastography in subgroup of 71 children with complete data and no documented exclusion criteria. Dots in b reflect individual participant stiffness measurements, box boundaries reflect first and third quartile values, and whiskers reflect 1.5 times the interquartile range.

Figure 2b:

(a) Histogram and (b) Tukey box/dot plot show distribution of liver stiffness values measured with MR elastography in subgroup of 71 children with complete data and no documented exclusion criteria. Dots in b reflect individual participant stiffness measurements, box boundaries reflect first and third quartile values, and whiskers reflect 1.5 times the interquartile range.

Figure 3:

Scatterplot of liver stiffness versus age for subgroup of 71 children with complete data and no documented exclusion criteria. There was no association between age and liver stiffness (r = 0.06; P = .62).